diff --git a/inst/tinytest/test-2-ipw-totals.R b/inst/tinytest/test-2-ipw-totals.R index bd4e25a..721e176 100644 --- a/inst/tinytest/test-2-ipw-totals.R +++ b/inst/tinytest/test-2-ipw-totals.R @@ -1,1276 +1,42 @@ -library(sampling) -library(survey) - - -# generation of data ---------------------------------------------------------------------- -set.seed(2024) -source("test-1-generate-data.R") ## path should be changed - - -# check logit ----------------------------------------------------------------------------- -## linear case ---------------------------------------------------------------------------- -#### correctly specified variables -------------------------------------------------------- -##### one target variable ---------------------------------------------------------------- - -## for y11 -expect_silent( - y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "logit") -) - -expect_equal(y11_corr_one$output$mean, 2.17757, tolerance = 0.0001) ## true value for this sim -expect_equal(y11_corr_one$output$SE, 0.1553234, tolerance = 0.0001) ## true value for this sim -expect_true(y11_corr_one$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_one$confidence_interval$upper_bound > mean(Y_11)) ## conf int - - -## for y12 -expect_silent( - y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_12, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "logit") -) -expect_equal(y12_corr_one$output$mean, 7.168049, tolerance = 0.0001) ## true value for this sim -expect_equal(y12_corr_one$output$SE, 1.01074, tolerance = 0.0001) ## true value for this sim -expect_true(y12_corr_one$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_one$confidence_interval$upper_bound > mean(Y_12)) ## conf int - - -## for y21 -expect_silent( - y21_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_21, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "logit") -) - -expect_equal(y21_corr_one$output$mean, 0.6675022, tolerance = 0.0001) ## true value for this sim -expect_equal(y21_corr_one$output$SE, 0.05836787, tolerance = 0.0001) ## true value for this sim -expect_true(y21_corr_one$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_one$confidence_interval$upper_bound > mean(Y_21)) ## conf int - -## for y22 -expect_silent( - y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_22, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "logit") -) - -expect_equal(y22_corr_one$output$mean, 0.6117923, tolerance = 0.0001) ## true value for this sim -expect_equal(y22_corr_one$output$SE, 0.06099473, tolerance = 0.0001) ## true value for this sim -expect_true(y22_corr_one$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_one$confidence_interval$upper_bound > mean(Y_22)) ## conf int - -##### all target variables --------------------------------------------------------------- - -expect_silent( - y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "logit") -) - -expect_identical(y_all_corr$output$mean, - c(y11_corr_one$output$mean, y12_corr_one$output$mean, - y21_corr_one$output$mean, y22_corr_one$output$mean)) - -expect_identical(y_all_corr$output$SE, - c(y11_corr_one$output$SE, y12_corr_one$output$SE, - y21_corr_one$output$SE, y22_corr_one$output$SE)) - -expect_identical(y_all_corr$confidence_interval, - data.frame(lower_bound = c(y11_corr_one$confidence_interval$lower_bound, - y12_corr_one$confidence_interval$lower_bound, - y21_corr_one$confidence_interval$lower_bound, - y22_corr_one$confidence_interval$lower_bound), - upper_bound = c(y11_corr_one$confidence_interval$upper_bound, - y12_corr_one$confidence_interval$upper_bound, - y21_corr_one$confidence_interval$upper_bound, - y22_corr_one$confidence_interval$upper_bound), - row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) - - - - - -#### all X variables variables ------------------------------------------------------------ -##### one target variable ---------------------------------------------------------------- - -## for y11 -expect_silent( - y11_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11, - data = sample_B1, - pop_totals = X_totals, - method_selection = "logit") -) - -expect_equal(y11_corr_all$output$mean, 2.005841, tolerance = 0.0001) ## true value for this sim -expect_equal(y11_corr_all$output$SE, 0.175094, tolerance = 0.0001) ## true value for this sim -expect_true(y11_corr_all$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_all$confidence_interval$upper_bound > mean(Y_11)) ## conf int - - -## for y12 -expect_silent( - y12_corr_all <- nonprob(selection = X_formula, - target = ~ Y_12, - data = sample_B1, - pop_totals = X_totals, - method_selection = "logit") -) -expect_equal(y12_corr_all$output$mean, 6.681308, tolerance = 0.0001) ## true value for this sim -expect_equal(y12_corr_all$output$SE, 0.7386469, tolerance = 0.0001) ## true value for this sim -expect_true(y12_corr_all$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_all$confidence_interval$upper_bound > mean(Y_12)) ## conf int - - -## for y21 -expect_silent( - y21_corr_all <- nonprob(selection = X_formula, - target = ~ Y_21, - data = sample_B1, - pop_totals = X_totals, - method_selection = "logit") -) - -expect_equal(y21_corr_all$output$mean, 0.64798484, tolerance = 0.0001) ## true value for this sim -expect_equal(y21_corr_all$output$SE, 0.061434896, tolerance = 0.0001) ## true value for this sim -expect_true(y21_corr_all$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_all$confidence_interval$upper_bound > mean(Y_21)) ## conf int - -## for y22 -expect_silent( - y22_corr_all <- nonprob(selection = X_formula, - target = ~ Y_22, - data = sample_B1, - pop_totals = X_totals, - method_selection = "logit") -) - -expect_equal(y22_corr_all$output$mean, 0.62883878, tolerance = 0.0001) ## true value for this sim -expect_equal(y22_corr_all$output$SE, 0.069170412, tolerance = 0.0001) ## true value for this sim -expect_true(y22_corr_all$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_all$confidence_interval$upper_bound > mean(Y_22)) ## conf int - -##### all target variables --------------------------------------------------------------- - -expect_silent( - y_all_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B1, - pop_totals = X_totals, - method_selection = "logit") -) - -expect_identical(y_all_corr_all$output$mean, - c(y11_corr_all$output$mean, y12_corr_all$output$mean, y21_corr_all$output$mean, y22_corr_all$output$mean)) - -expect_identical(y_all_corr_all$output$SE, - c(y11_corr_all$output$SE, y12_corr_all$output$SE, y21_corr_all$output$SE, y22_corr_all$output$SE)) - -expect_identical(y_all_corr_all$confidence_interval, - data.frame(lower_bound = c(y11_corr_all$confidence_interval$lower_bound, - y12_corr_all$confidence_interval$lower_bound, - y21_corr_all$confidence_interval$lower_bound, - y22_corr_all$confidence_interval$lower_bound), - upper_bound = c(y11_corr_all$confidence_interval$upper_bound, - y12_corr_all$confidence_interval$upper_bound, - y21_corr_all$confidence_interval$upper_bound, - y22_corr_all$confidence_interval$upper_bound), - row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) - -# These tests are only supposed to be run on developer's machine and -# package GitHub page not on CRAN (they take too long) - -if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { - #### variable selection ------------------------------------------------------------------ - ##### one target variable ---------------------------------------------------------------- - - ## y_11 - expect_silent( - y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_11, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "logit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y11_corr_scad$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim - expect_equal(y11_corr_scad$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim - expect_false(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int - expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) - - ## y_12 - expect_silent( - y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_12, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "logit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y12_corr_scad$output$mean, 6.9530644, tolerance = 0.0001) ## true value for this sim - expect_equal(y12_corr_scad$output$SE, 0.15341599, tolerance = 0.0001) ## true value for this sim - expect_true(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int - expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) - - ## y_21 - expect_silent( - y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_21, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "logit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y21_corr_scad$output$mean, 0.78264707, tolerance = 0.0001) ## true value for this sim - expect_equal(y21_corr_scad$output$SE, 0.0090012565, tolerance = 0.0001) ## true value for this sim - expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int - expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) - - ## y_22 - expect_silent( - y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_22, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "logit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y22_corr_scad$output$mean, 0.57680653, tolerance = 0.0001) ## true value for this sim - expect_equal(y22_corr_scad$output$SE, 0.011240221, tolerance = 0.0001) ## true value for this sim - expect_false(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int - expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) - - - ## lasso - - # expect_silent( - # y11_corr_lasso <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "logit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "lasso")) - # ) - # - # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim - # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim - # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & - # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int - # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) - - ## MCP - - # expect_silent( - # y11_corr_mcp <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "logit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "MCP")) - # ) - # - # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim - # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim - # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & - # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int - # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) - - ##### all target variables --------------------------------------------------------------- - - # expect_silent( - # y_all_corr_all <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, + Y_12 + Y_21 + Y_22, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "logit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "SCAD", nfolds = 5), - # verbose = T) - # ) - -} - - -## non-linear case ------------------------------------------------------------------------ -#### correctly specified variables -------------------------------------------------------- -##### one target variable ---------------------------------------------------------------- - -## for y11 -expect_silent( - y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "logit") -) - -expect_equal(y11_corr_one$output$mean, 1.9842061, tolerance = 0.0001) ## true value for this sim -expect_equal(y11_corr_one$output$SE, 0.1141722, tolerance = 0.0001) ## true value for this sim -expect_true(y11_corr_one$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_one$confidence_interval$upper_bound > mean(Y_11)) ## conf int - - -## for y12 -expect_silent( - y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_12, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "logit") -) -expect_equal(y12_corr_one$output$mean, 5.7852985, tolerance = 0.0001) ## true value for this sim -expect_equal(y12_corr_one$output$SE, 0.25978825, tolerance = 0.0001) ## true value for this sim -expect_false(y12_corr_one$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_one$confidence_interval$upper_bound > mean(Y_12)) ## conf int - - -## for y21 -expect_silent( - y21_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_21, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "logit") -) - -expect_equal(y21_corr_one$output$mean, 0.61846098, tolerance = 0.0001) ## true value for this sim -expect_equal(y21_corr_one$output$SE, 0.026561368, tolerance = 0.0001) ## true value for this sim -expect_true(y21_corr_one$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_one$confidence_interval$upper_bound > mean(Y_21)) ## conf int - -## for y22 -expect_silent( - y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_22, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "logit") -) - -expect_equal(y22_corr_one$output$mean, 0.64512141, tolerance = 0.0001) ## true value for this sim -expect_equal(y22_corr_one$output$SE, 0.026011784, tolerance = 0.0001) ## true value for this sim -expect_true(y22_corr_one$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_one$confidence_interval$upper_bound > mean(Y_22)) ## conf int - -##### all target variables --------------------------------------------------------------- - -expect_silent( - y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "logit") -) - -expect_identical(y_all_corr$output$mean, - c(y11_corr_one$output$mean, y12_corr_one$output$mean, - y21_corr_one$output$mean, y22_corr_one$output$mean)) - -expect_identical(y_all_corr$output$SE, - c(y11_corr_one$output$SE, y12_corr_one$output$SE, - y21_corr_one$output$SE, y22_corr_one$output$SE)) - -expect_identical(y_all_corr$confidence_interval, - data.frame(lower_bound = c(y11_corr_one$confidence_interval$lower_bound, - y12_corr_one$confidence_interval$lower_bound, - y21_corr_one$confidence_interval$lower_bound, - y22_corr_one$confidence_interval$lower_bound), - upper_bound = c(y11_corr_one$confidence_interval$upper_bound, - y12_corr_one$confidence_interval$upper_bound, - y21_corr_one$confidence_interval$upper_bound, - y22_corr_one$confidence_interval$upper_bound), - row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) - - - - - -#### all X variables variables ------------------------------------------------------------ -##### one target variable ---------------------------------------------------------------- - -## for y11 -expect_silent( - y11_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11, - data = sample_B2, - pop_totals = X_totals, - method_selection = "logit") -) - -expect_equal(y11_corr_all$output$mean, 1.9820754, tolerance = 0.0001) ## true value for this sim -expect_equal(y11_corr_all$output$SE, 0.13574794, tolerance = 0.0001) ## true value for this sim -expect_true(y11_corr_all$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_all$confidence_interval$upper_bound > mean(Y_11)) ## conf int - - -## for y12 -expect_silent( - y12_corr_all <- nonprob(selection = X_formula, - target = ~ Y_12, - data = sample_B2, - pop_totals = X_totals, - method_selection = "logit") -) -expect_equal(y12_corr_all$output$mean, 5.6417776, tolerance = 0.0001) ## true value for this sim -expect_equal(y12_corr_all$output$SE, 0.27204781, tolerance = 0.0001) ## true value for this sim -expect_false(y12_corr_all$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_all$confidence_interval$upper_bound > mean(Y_12)) ## conf int - - -## for y21 -expect_silent( - y21_corr_all <- nonprob(selection = X_formula, - target = ~ Y_21, - data = sample_B2, - pop_totals = X_totals, - method_selection = "logit") -) - -expect_equal(y21_corr_all$output$mean, 0.61556545, tolerance = 0.0001) ## true value for this sim -expect_equal(y21_corr_all$output$SE, 0.028680898, tolerance = 0.0001) ## true value for this sim -expect_true(y21_corr_all$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_all$confidence_interval$upper_bound > mean(Y_21)) ## conf int - -## for y22 -expect_silent( - y22_corr_all <- nonprob(selection = X_formula, - target = ~ Y_22, - data = sample_B2, - pop_totals = X_totals, - method_selection = "logit") -) - -expect_equal(y22_corr_all$output$mean, 0.57173102, tolerance = 0.0001) ## true value for this sim -expect_equal(y22_corr_all$output$SE, 0.024335536, tolerance = 0.0001) ## true value for this sim -expect_false(y22_corr_all$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_all$confidence_interval$upper_bound > mean(Y_22)) ## conf int - -##### all target variables --------------------------------------------------------------- - -expect_silent( - y_all_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B2, - pop_totals = X_totals, - method_selection = "logit") -) - -expect_identical(y_all_corr_all$output$mean, - c(y11_corr_all$output$mean, y12_corr_all$output$mean, y21_corr_all$output$mean, y22_corr_all$output$mean)) - -expect_identical(y_all_corr_all$output$SE, - c(y11_corr_all$output$SE, y12_corr_all$output$SE, y21_corr_all$output$SE, y22_corr_all$output$SE)) - -expect_identical(y_all_corr_all$confidence_interval, - data.frame(lower_bound = c(y11_corr_all$confidence_interval$lower_bound, - y12_corr_all$confidence_interval$lower_bound, - y21_corr_all$confidence_interval$lower_bound, - y22_corr_all$confidence_interval$lower_bound), - upper_bound = c(y11_corr_all$confidence_interval$upper_bound, - y12_corr_all$confidence_interval$upper_bound, - y21_corr_all$confidence_interval$upper_bound, - y22_corr_all$confidence_interval$upper_bound), - row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) - - - -#### variable selection ------------------------------------------------------------------ -##### one target variable ---------------------------------------------------------------- - -if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { - - ## y_11 - expect_silent( - y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_11, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "logit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y11_corr_scad$output$mean, 1.8810431, tolerance = 0.0001) ## true value for this sim - expect_equal(y11_corr_scad$output$SE, 0.059381217, tolerance = 0.0001) ## true value for this sim - expect_true(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int - expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) - - ## y_12 - expect_silent( - y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_12, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "logit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y12_corr_scad$output$mean, 5.796713, tolerance = 0.0001) ## true value for this sim - expect_equal(y12_corr_scad$output$SE, 0.14583111, tolerance = 0.0001) ## true value for this sim - expect_false(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int - expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) - - # y_21 - expect_silent( - y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_21, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "logit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y21_corr_scad$output$mean, 0.6060074, tolerance = 0.0001) ## true value for this sim - expect_equal(y21_corr_scad$output$SE, 0.010194911, tolerance = 0.0001) ## true value for this sim - expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int - expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) - - # # y_22 - expect_silent( - y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_22, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "logit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y22_corr_scad$output$mean, 0.64707641, tolerance = 0.0001) ## true value for this sim - expect_equal(y22_corr_scad$output$SE, 0.0099648982, tolerance = 0.0001) ## true value for this sim - expect_true(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int - expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) - - - ## lasso - - # expect_silent( - # y11_corr_lasso <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "logit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "lasso")) - # ) - # - # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim - # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim - # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & - # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int - # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) - - ## MCP - - # expect_silent( - # y11_corr_mcp <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "logit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "MCP")) - # ) - # - # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim - # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim - # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & - # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int - # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) - - ##### all target variables --------------------------------------------------------------- - - # expect_silent( - # y_all_corr_all <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, + Y_12 + Y_21 + Y_22, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "logit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "SCAD", nfolds = 5), - # verbose = T) - # ) - -} - - -# check probit ---------------------------------------------------------------------------- -## linear case ---------------------------------------------------------------------------- -#### correctly specified variables -------------------------------------------------------- -##### one target variable ---------------------------------------------------------------- - -## for y11 -expect_silent( - y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "probit") -) - -expect_equal(y11_corr_one$output$mean, 2.1985815, tolerance = 0.0001) ## true value for this sim -expect_equal(y11_corr_one$output$SE, 0.12852317, tolerance = 0.0001) ## true value for this sim -expect_true(y11_corr_one$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_one$confidence_interval$upper_bound > mean(Y_11)) ## conf int - - -## for y12 -expect_silent( - y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_12, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "probit") -) -expect_equal(y12_corr_one$output$mean, 7.1226154, tolerance = 0.0001) ## true value for this sim -expect_equal(y12_corr_one$output$SE, 0.83353647, tolerance = 0.0001) ## true value for this sim -expect_true(y12_corr_one$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_one$confidence_interval$upper_bound > mean(Y_12)) ## conf int - - -## for y21 -expect_silent( - y21_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_21, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "probit") -) - -expect_equal(y21_corr_one$output$mean, 0.65778764, tolerance = 0.0001) ## true value for this sim -expect_equal(y21_corr_one$output$SE, 0.066138969, tolerance = 0.0001) ## true value for this sim -expect_true(y21_corr_one$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_one$confidence_interval$upper_bound > mean(Y_21)) ## conf int - -## for y22 -expect_silent( - y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_22, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "probit") -) - -expect_equal(y22_corr_one$output$mean, 0.58148029, tolerance = 0.0001) ## true value for this sim -expect_equal(y22_corr_one$output$SE, 0.068800601, tolerance = 0.0001) ## true value for this sim -expect_true(y22_corr_one$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_one$confidence_interval$upper_bound > mean(Y_22)) ## conf int - -##### all target variables --------------------------------------------------------------- - -expect_silent( - y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "probit") -) - -expect_identical(y_all_corr$output$mean, - c(y11_corr_one$output$mean, y12_corr_one$output$mean, - y21_corr_one$output$mean, y22_corr_one$output$mean)) - -expect_identical(y_all_corr$output$SE, - c(y11_corr_one$output$SE, y12_corr_one$output$SE, - y21_corr_one$output$SE, y22_corr_one$output$SE)) - -expect_identical(y_all_corr$confidence_interval, - data.frame(lower_bound = c(y11_corr_one$confidence_interval$lower_bound, - y12_corr_one$confidence_interval$lower_bound, - y21_corr_one$confidence_interval$lower_bound, - y22_corr_one$confidence_interval$lower_bound), - upper_bound = c(y11_corr_one$confidence_interval$upper_bound, - y12_corr_one$confidence_interval$upper_bound, - y21_corr_one$confidence_interval$upper_bound, - y22_corr_one$confidence_interval$upper_bound), - row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) - - - - - -#### all X variables variables ------------------------------------------------------------ -##### one target variable ---------------------------------------------------------------- - -## for y11 -expect_silent( - y11_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11, - data = sample_B1, - pop_totals = X_totals, - method_selection = "probit") -) - -expect_equal(y11_corr_all$output$mean, 2.0102108, tolerance = 0.0001) ## true value for this sim -expect_equal(y11_corr_all$output$SE, 0.16422236, tolerance = 0.0001) ## true value for this sim -expect_true(y11_corr_all$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_all$confidence_interval$upper_bound > mean(Y_11)) ## conf int - - -## for y12 -expect_silent( - y12_corr_all <- nonprob(selection = X_formula, - target = ~ Y_12, - data = sample_B1, - pop_totals = X_totals, - method_selection = "probit") -) -expect_equal(y12_corr_all$output$mean, 6.6878138, tolerance = 0.0001) ## true value for this sim -expect_equal(y12_corr_all$output$SE, 0.60740161, tolerance = 0.0001) ## true value for this sim -expect_true(y12_corr_all$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_all$confidence_interval$upper_bound > mean(Y_12)) ## conf int - - -## for y21 -expect_silent( - y21_corr_all <- nonprob(selection = X_formula, - target = ~ Y_21, - data = sample_B1, - pop_totals = X_totals, - method_selection = "probit") -) - -expect_equal(y21_corr_all$output$mean, 0.64617883, tolerance = 0.0001) ## true value for this sim -expect_equal(y21_corr_all$output$SE, 0.074037561, tolerance = 0.0001) ## true value for this sim -expect_true(y21_corr_all$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_all$confidence_interval$upper_bound > mean(Y_21)) ## conf int - -## for y22 -expect_silent( - y22_corr_all <- nonprob(selection = X_formula, - target = ~ Y_22, - data = sample_B1, - pop_totals = X_totals, - method_selection = "probit") -) - -expect_equal(y22_corr_all$output$mean, 0.62456575, tolerance = 0.0001) ## true value for this sim -expect_equal(y22_corr_all$output$SE, 0.083014906, tolerance = 0.0001) ## true value for this sim -expect_true(y22_corr_all$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_all$confidence_interval$upper_bound > mean(Y_22)) ## conf int - -##### all target variables --------------------------------------------------------------- - -expect_silent( - y_all_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B1, - pop_totals = X_totals, - method_selection = "probit") -) - -expect_identical(y_all_corr_all$output$mean, - c(y11_corr_all$output$mean, y12_corr_all$output$mean, y21_corr_all$output$mean, y22_corr_all$output$mean)) - -expect_identical(y_all_corr_all$output$SE, - c(y11_corr_all$output$SE, y12_corr_all$output$SE, y21_corr_all$output$SE, y22_corr_all$output$SE)) - -expect_identical(y_all_corr_all$confidence_interval, - data.frame(lower_bound = c(y11_corr_all$confidence_interval$lower_bound, - y12_corr_all$confidence_interval$lower_bound, - y21_corr_all$confidence_interval$lower_bound, - y22_corr_all$confidence_interval$lower_bound), - upper_bound = c(y11_corr_all$confidence_interval$upper_bound, - y12_corr_all$confidence_interval$upper_bound, - y21_corr_all$confidence_interval$upper_bound, - y22_corr_all$confidence_interval$upper_bound), - row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) - - -if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { - #### variable selection ------------------------------------------------------------------ - ##### one target variable ---------------------------------------------------------------- - - ## y_11 - expect_silent( - y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_11, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "probit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y11_corr_scad$output$mean, 3.0633399, tolerance = 0.0001) ## true value for this sim - expect_equal(y11_corr_scad$output$SE, 0.049384577, tolerance = 0.0001) ## true value for this sim - expect_false(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int - expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) - - ## y_12 - expect_silent( - y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_12, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "probit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y12_corr_scad$output$mean, 6.9420676, tolerance = 0.0001) ## true value for this sim - expect_equal(y12_corr_scad$output$SE, 0.15578019, tolerance = 0.0001) ## true value for this sim - expect_true(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int - expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) - - ## y_21 - expect_silent( - y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_21, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "probit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y21_corr_scad$output$mean, 0.78324543, tolerance = 0.0001) ## true value for this sim - expect_equal(y21_corr_scad$output$SE, 0.0090654702, tolerance = 0.0001) ## true value for this sim - expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int - expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) - - ## y_22 - expect_silent( - y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_22, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "probit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y22_corr_scad$output$mean, 0.57672297, tolerance = 0.0001) ## true value for this sim - expect_equal(y22_corr_scad$output$SE, 0.011433876, tolerance = 0.0001) ## true value for this sim - expect_false(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int - expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) - - - ## lasso - - # expect_silent( - # y11_corr_lasso <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "probit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "lasso")) - # ) - # - # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim - # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim - # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & - # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int - # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) - - ## MCP - - # expect_silent( - # y11_corr_mcp <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "probit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "MCP")) - # ) - # - # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim - # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim - # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & - # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int - # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) - - ##### all target variables --------------------------------------------------------------- - - # expect_silent( - # y_all_corr_all <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, + Y_12 + Y_21 + Y_22, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "probit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "SCAD", nfolds = 5), - # verbose = T) - # ) - -} - -## non-linear case ------------------------------------------------------------------------ -#### correctly specified variables -------------------------------------------------------- -##### one target variable ---------------------------------------------------------------- - -## for y11 -expect_silent( - y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "probit") -) - -expect_equal(y11_corr_one$output$mean, 1.984232, tolerance = 0.0001) ## true value for this sim -expect_equal(y11_corr_one$output$SE, 0.10478658, tolerance = 0.0001) ## true value for this sim -expect_true(y11_corr_one$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_one$confidence_interval$upper_bound > mean(Y_11)) ## conf int - - -## for y12 -expect_silent( - y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_12, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "probit") -) -expect_equal(y12_corr_one$output$mean, 5.7851563, tolerance = 0.0001) ## true value for this sim -expect_equal(y12_corr_one$output$SE, 0.20174126, tolerance = 0.0001) ## true value for this sim -expect_false(y12_corr_one$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_one$confidence_interval$upper_bound > mean(Y_12)) ## conf int - - -## for y21 -expect_silent( - y21_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_21, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "probit") -) - -expect_equal(y21_corr_one$output$mean, 0.61846032, tolerance = 0.0001) ## true value for this sim -expect_equal(y21_corr_one$output$SE, 0.031103296, tolerance = 0.0001) ## true value for this sim -expect_true(y21_corr_one$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_one$confidence_interval$upper_bound > mean(Y_21)) ## conf int - -## for y22 -expect_silent( - y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_22, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "probit") -) - -expect_equal(y22_corr_one$output$mean, 0.64509743, tolerance = 0.0001) ## true value for this sim -expect_equal(y22_corr_one$output$SE, 0.030385359, tolerance = 0.0001) ## true value for this sim -expect_true(y22_corr_one$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_one$confidence_interval$upper_bound > mean(Y_22)) ## conf int - -##### all target variables --------------------------------------------------------------- - -expect_silent( - y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "probit") -) - -expect_identical(y_all_corr$output$mean, - c(y11_corr_one$output$mean, y12_corr_one$output$mean, - y21_corr_one$output$mean, y22_corr_one$output$mean)) - -expect_identical(y_all_corr$output$SE, - c(y11_corr_one$output$SE, y12_corr_one$output$SE, - y21_corr_one$output$SE, y22_corr_one$output$SE)) - -expect_identical(y_all_corr$confidence_interval, - data.frame(lower_bound = c(y11_corr_one$confidence_interval$lower_bound, - y12_corr_one$confidence_interval$lower_bound, - y21_corr_one$confidence_interval$lower_bound, - y22_corr_one$confidence_interval$lower_bound), - upper_bound = c(y11_corr_one$confidence_interval$upper_bound, - y12_corr_one$confidence_interval$upper_bound, - y21_corr_one$confidence_interval$upper_bound, - y22_corr_one$confidence_interval$upper_bound), - row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) - - - - - -#### all X variables variables ------------------------------------------------------------ -##### one target variable ---------------------------------------------------------------- - -## for y11 -expect_silent( - y11_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11, - data = sample_B2, - pop_totals = X_totals, - method_selection = "probit") -) - -expect_equal(y11_corr_all$output$mean, 1.981833, tolerance = 0.0001) ## true value for this sim -expect_equal(y11_corr_all$output$SE, 0.12730647, tolerance = 0.0001) ## true value for this sim -expect_true(y11_corr_all$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_all$confidence_interval$upper_bound > mean(Y_11)) ## conf int - - -## for y12 -expect_silent( - y12_corr_all <- nonprob(selection = X_formula, - target = ~ Y_12, - data = sample_B2, - pop_totals = X_totals, - method_selection = "probit") -) -expect_equal(y12_corr_all$output$mean, 5.6436133, tolerance = 0.0001) ## true value for this sim -expect_equal(y12_corr_all$output$SE, 0.21772187, tolerance = 0.0001) ## true value for this sim -expect_false(y12_corr_all$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_all$confidence_interval$upper_bound > mean(Y_12)) ## conf int - - -## for y21 -expect_silent( - y21_corr_all <- nonprob(selection = X_formula, - target = ~ Y_21, - data = sample_B2, - pop_totals = X_totals, - method_selection = "probit") -) - -expect_equal(y21_corr_all$output$mean, 0.61562293, tolerance = 0.0001) ## true value for this sim -expect_equal(y21_corr_all$output$SE, 0.033259168, tolerance = 0.0001) ## true value for this sim -expect_true(y21_corr_all$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_all$confidence_interval$upper_bound > mean(Y_21)) ## conf int - -## for y22 -expect_silent( - y22_corr_all <- nonprob(selection = X_formula, - target = ~ Y_22, - data = sample_B2, - pop_totals = X_totals, - method_selection = "probit") -) - -expect_equal(y22_corr_all$output$mean, 0.57173464, tolerance = 0.0001) ## true value for this sim -expect_equal(y22_corr_all$output$SE, 0.028962455, tolerance = 0.0001) ## true value for this sim -expect_false(y22_corr_all$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_all$confidence_interval$upper_bound > mean(Y_22)) ## conf int - -##### all target variables --------------------------------------------------------------- - -expect_silent( - y_all_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B2, - pop_totals = X_totals, - method_selection = "probit") -) - -expect_identical(y_all_corr_all$output$mean, - c(y11_corr_all$output$mean, y12_corr_all$output$mean, - y21_corr_all$output$mean, y22_corr_all$output$mean)) - -expect_identical(y_all_corr_all$output$SE, - c(y11_corr_all$output$SE, y12_corr_all$output$SE, - y21_corr_all$output$SE, y22_corr_all$output$SE)) - -expect_identical(y_all_corr_all$confidence_interval, - data.frame(lower_bound = c(y11_corr_all$confidence_interval$lower_bound, - y12_corr_all$confidence_interval$lower_bound, - y21_corr_all$confidence_interval$lower_bound, - y22_corr_all$confidence_interval$lower_bound), - upper_bound = c(y11_corr_all$confidence_interval$upper_bound, - y12_corr_all$confidence_interval$upper_bound, - y21_corr_all$confidence_interval$upper_bound, - y22_corr_all$confidence_interval$upper_bound), - row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) - - -if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { - #### variable selection ------------------------------------------------------------------ - ##### one target variable ---------------------------------------------------------------- - - # # y_11 - expect_silent( - y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_11, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "probit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y11_corr_scad$output$mean, 1.8810431, tolerance = 0.0001) ## true value for this sim - expect_equal(y11_corr_scad$output$SE, 0.059381198, tolerance = 0.0001) ## true value for this sim - expect_true(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & - y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int - expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) - - ## y_12 - expect_silent( - y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_12, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "probit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y12_corr_scad$output$mean, 5.7967136, tolerance = 0.0001) ## true value for this sim - expect_equal(y12_corr_scad$output$SE, 0.14583128, tolerance = 0.0001) ## true value for this sim - expect_false(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & - y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int - expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) - - # # y_21 - expect_silent( - y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_21, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "probit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y21_corr_scad$output$mean, 0.60600756, tolerance = 0.0001) ## true value for this sim - expect_equal(y21_corr_scad$output$SE, 0.010194928, tolerance = 0.0001) ## true value for this sim - expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & - y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int - expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) - - # # y_22 - expect_silent( - y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_22, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "probit", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - expect_equal(y22_corr_scad$output$mean, 0.64707626, tolerance = 0.0001) ## true value for this sim - expect_equal(y22_corr_scad$output$SE, 0.0099648984, tolerance = 0.0001) ## true value for this sim - expect_true(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & - y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int - expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) - - - ## lasso - - # expect_silent( - # y11_corr_lasso <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "probit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "lasso")) - # ) - # - # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim - # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim - # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & - # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int - # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) - - ## MCP - - # expect_silent( - # y11_corr_mcp <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "probit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "MCP")) - # ) - # - # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim - # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim - # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & - # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int - # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) - - ##### all target variables --------------------------------------------------------------- - - # expect_silent( - # y_all_corr_all <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - # target = ~ Y_11, + Y_12 + Y_21 + Y_22, - # data = sample_B1, - # pop_totals = X_totals[1:11], - # method_selection = "probit", - # control_inference = controlInf(vars_selection = TRUE), - # control_selection = controlSel(penalty = "SCAD", nfolds = 5), - # verbose = T) - # ) - -} - -# check cloglog ----------------------------------------------------- -## linear case ---------------------------------------------------------------------------- -#### correctly specified variables -------------------------------------------------------- -##### one target variable ---------------------------------------------------------------- - -## for y11 -expect_silent( - y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "cloglog") -) -# -# expect_equal(y11_corr_one$output$mean, 2.015028, tolerance = 0.0001) ## true value for this sim -# expect_equal(y11_corr_one$output$SE, 0.1713938, tolerance = 0.0001) ## true value for this sim +# library(sampling) +# library(survey) +# +# +# # generation of data ---------------------------------------------------------------------- +# set.seed(2024) +# source("test-1-generate-data.R") ## path should be changed +# +# +# # check logit ----------------------------------------------------------------------------- +# ## linear case ---------------------------------------------------------------------------- +# #### correctly specified variables -------------------------------------------------------- +# ##### one target variable ---------------------------------------------------------------- +# +# ## for y11 +# expect_silent( +# y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "logit") +# ) +# +# expect_equal(y11_corr_one$output$mean, 2.17757, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_one$output$SE, 0.1553234, tolerance = 0.0001) ## true value for this sim # expect_true(y11_corr_one$confidence_interval$lower_bound < mean(Y_11) & # y11_corr_one$confidence_interval$upper_bound > mean(Y_11)) ## conf int - - +# +# # ## for y12 -expect_silent( - y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_12, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "cloglog") -) -# expect_equal(y12_corr_one$output$mean, 6.617158, tolerance = 0.0001) ## true value for this sim -# expect_equal(y12_corr_one$output$SE, 0.6183295, tolerance = 0.0001) ## true value for this sim +# expect_silent( +# y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_12, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "logit") +# ) +# expect_equal(y12_corr_one$output$mean, 7.168049, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_one$output$SE, 1.01074, tolerance = 0.0001) ## true value for this sim # expect_true(y12_corr_one$confidence_interval$lower_bound < mean(Y_12) & # y12_corr_one$confidence_interval$upper_bound > mean(Y_12)) ## conf int # @@ -1278,40 +44,40 @@ expect_silent( # ## for y21 # expect_silent( # y21_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, -# target = ~ Y_21, -# data = sample_B1, -# pop_totals = X_totals[1:5], -# method_selection = "cloglog") +# target = ~ Y_21, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "logit") # ) # -# expect_equal(y21_corr_one$output$mean, 0.6589795, tolerance = 0.0001) ## true value for this sim -# expect_equal(y21_corr_one$output$SE, 0.05471915, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_one$output$mean, 0.6675022, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_one$output$SE, 0.05836787, tolerance = 0.0001) ## true value for this sim # expect_true(y21_corr_one$confidence_interval$lower_bound < mean(Y_21) & # y21_corr_one$confidence_interval$upper_bound > mean(Y_21)) ## conf int # # ## for y22 -expect_silent( - y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_22, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "cloglog") -) -# -# expect_equal(y22_corr_one$output$mean, 0.6872036, tolerance = 0.0001) ## true value for this sim -# expect_equal(y22_corr_one$output$SE, 0.06540456, tolerance = 0.0001) ## true value for this sim +# expect_silent( +# y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_22, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "logit") +# ) +# +# expect_equal(y22_corr_one$output$mean, 0.6117923, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_one$output$SE, 0.06099473, tolerance = 0.0001) ## true value for this sim # expect_true(y22_corr_one$confidence_interval$lower_bound < mean(Y_22) & # y22_corr_one$confidence_interval$upper_bound > mean(Y_22)) ## conf int # # ##### all target variables --------------------------------------------------------------- # -expect_silent( - y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B1, - pop_totals = X_totals[1:5], - method_selection = "cloglog") -) +# expect_silent( +# y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "logit") +# ) # # expect_identical(y_all_corr$output$mean, # c(y11_corr_one$output$mean, y12_corr_one$output$mean, @@ -1320,7 +86,7 @@ expect_silent( # expect_identical(y_all_corr$output$SE, # c(y11_corr_one$output$SE, y12_corr_one$output$SE, # y21_corr_one$output$SE, y22_corr_one$output$SE)) -# # +# # expect_identical(y_all_corr$confidence_interval, # data.frame(lower_bound = c(y11_corr_one$confidence_interval$lower_bound, # y12_corr_one$confidence_interval$lower_bound, @@ -1334,75 +100,77 @@ expect_silent( # # # -# #### all X variables variables TOFIX ------------------------------------------------------------ +# +# +# #### all X variables variables ------------------------------------------------------------ # ##### one target variable ---------------------------------------------------------------- # # ## for y11 -expect_silent( - y11_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11, - data = sample_B1, - pop_totals = X_totals, - method_selection = "cloglog") -) -# -# expect_equal(y11_corr_all$output$mean, 1.956418, tolerance = 0.0001) ## true value for this sim -# expect_equal(y11_corr_all$output$SE, 0.2019104, tolerance = 0.0001) ## true value for this sim +# expect_silent( +# y11_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "logit") +# ) +# +# expect_equal(y11_corr_all$output$mean, 2.005841, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_all$output$SE, 0.175094, tolerance = 0.0001) ## true value for this sim # expect_true(y11_corr_all$confidence_interval$lower_bound < mean(Y_11) & # y11_corr_all$confidence_interval$upper_bound > mean(Y_11)) ## conf int # # # ## for y12 -expect_silent( - y12_corr_all <- nonprob(selection = X_formula, - target = ~ Y_12, - data = sample_B1, - pop_totals = X_totals, - method_selection = "cloglog") -) -# expect_equal(y12_corr_all$output$mean, 6.571494, tolerance = 0.0001) ## true value for this sim -# expect_equal(y12_corr_all$output$SE, 0.7128783, tolerance = 0.0001) ## true value for this sim +# expect_silent( +# y12_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_12, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "logit") +# ) +# expect_equal(y12_corr_all$output$mean, 6.681308, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_all$output$SE, 0.7386469, tolerance = 0.0001) ## true value for this sim # expect_true(y12_corr_all$confidence_interval$lower_bound < mean(Y_12) & # y12_corr_all$confidence_interval$upper_bound > mean(Y_12)) ## conf int # # # ## for y21 -expect_silent( - y21_corr_all <- nonprob(selection = X_formula, - target = ~ Y_21, - data = sample_B1, - pop_totals = X_totals, - method_selection = "cloglog") -) -# -# expect_equal(y21_corr_all$output$mean, 0.6496145, tolerance = 0.0001) ## true value for this sim -# expect_equal(y21_corr_all$output$SE, 0.0575112, tolerance = 0.0001) ## true value for this sim +# expect_silent( +# y21_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_21, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "logit") +# ) +# +# expect_equal(y21_corr_all$output$mean, 0.64798484, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_all$output$SE, 0.061434896, tolerance = 0.0001) ## true value for this sim # expect_true(y21_corr_all$confidence_interval$lower_bound < mean(Y_21) & # y21_corr_all$confidence_interval$upper_bound > mean(Y_21)) ## conf int # # ## for y22 -expect_silent( - y22_corr_all <- nonprob(selection = X_formula, - target = ~ Y_22, - data = sample_B1, - pop_totals = X_totals, - method_selection = "cloglog") -) -# -# expect_equal(y22_corr_all$output$mean, 0.6763156, tolerance = 0.0001) ## true value for this sim -# expect_equal(y22_corr_all$output$SE, 0.06510848, tolerance = 0.0001) ## true value for this sim +# expect_silent( +# y22_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_22, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "logit") +# ) +# +# expect_equal(y22_corr_all$output$mean, 0.62883878, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_all$output$SE, 0.069170412, tolerance = 0.0001) ## true value for this sim # expect_true(y22_corr_all$confidence_interval$lower_bound < mean(Y_22) & # y22_corr_all$confidence_interval$upper_bound > mean(Y_22)) ## conf int # # ##### all target variables --------------------------------------------------------------- # -expect_silent( - y_all_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B1, - pop_totals = X_totals, - method_selection = "cloglog") -) +# expect_silent( +# y_all_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "logit") +# ) # # expect_identical(y_all_corr_all$output$mean, # c(y11_corr_all$output$mean, y12_corr_all$output$mean, y21_corr_all$output$mean, y22_corr_all$output$mean)) @@ -1421,153 +189,204 @@ expect_silent( # y22_corr_all$confidence_interval$upper_bound), # row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) # - -if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { - - #### variable selection ------------------------------------------------------------------ - ##### one target variable ---------------------------------------------------------------- - - ## y_11 to fix - expect_silent( - y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_11, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "cloglog", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - # expect_equal(y11_corr_scad$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim - # expect_equal(y11_corr_scad$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim - # expect_false(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & - # y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int - # expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) - - ## y_12 to fix - expect_silent( - y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_12, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "cloglog", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - # expect_equal(y12_corr_scad$output$mean, 6.9530644, tolerance = 0.0001) ## true value for this sim - # expect_equal(y12_corr_scad$output$SE, 0.15341599, tolerance = 0.0001) ## true value for this sim - # expect_true(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & - # y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int - # expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) - - ## y_21 to fix - expect_silent( - y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_21, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "cloglog", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - # expect_equal(y21_corr_scad$output$mean, 0.78264707, tolerance = 0.0001) ## true value for this sim - # expect_equal(y21_corr_scad$output$SE, 0.0090012565, tolerance = 0.0001) ## true value for this sim - # expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & - # y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int - # expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) - - ## y_22 to fix - expect_silent( - y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_22, - data = sample_B1, - pop_totals = X_totals[1:11], - method_selection = "cloglog", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - - # expect_equal(y22_corr_scad$output$mean, 0.57680653, tolerance = 0.0001) ## true value for this sim - # expect_equal(y22_corr_scad$output$SE, 0.011240221, tolerance = 0.0001) ## true value for this sim - # expect_false(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & - # y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int - # expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) - - - ## non-linear case ------------------------------------------------------------------------ - #### correctly specified variables -------------------------------------------------------- - ##### one target variable ---------------------------------------------------------------- - - ## for y11 - expect_silent( - y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "cloglog") - ) - # - # expect_equal(y11_corr_one$output$mean, 2.10861, tolerance = 0.0001) ## true value for this sim - # expect_equal(y11_corr_one$output$SE, 0.1243684, tolerance = 0.0001) ## true value for this sim - # expect_true(y11_corr_one$confidence_interval$lower_bound < mean(Y_11) & - # y11_corr_one$confidence_interval$upper_bound > mean(Y_11)) ## conf int - # - # - # ## for y12 - expect_silent( - y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_12, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "cloglog") - ) - # expect_equal(y12_corr_one$output$mean, 5.708058, tolerance = 0.0001) ## true value for this sim - # expect_equal(y12_corr_one$output$SE, 0.2604017, tolerance = 0.0001) ## true value for this sim - # expect_false(y12_corr_one$confidence_interval$lower_bound < mean(Y_12) & - # y12_corr_one$confidence_interval$upper_bound > mean(Y_12)) ## conf int - # - # - # ## for y21 - expect_silent( - y21_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_21, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "cloglog") - ) - # - # expect_equal(y21_corr_one$output$mean, 0.6096105, tolerance = 0.0001) ## true value for this sim - # expect_equal(y21_corr_one$output$SE, 0.02698877, tolerance = 0.0001) ## true value for this sim - # expect_true(y21_corr_one$confidence_interval$lower_bound < mean(Y_21) & - # y21_corr_one$confidence_interval$upper_bound > mean(Y_21)) ## conf int - # - # ## for y22 - expect_silent( - y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_22, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "cloglog") - ) - # - # expect_equal(y22_corr_one$output$mean, 0.6210885, tolerance = 0.0001) ## true value for this sim - # expect_equal(y22_corr_one$output$SE, 0.02561999, tolerance = 0.0001) ## true value for this sim - # expect_true(y22_corr_one$confidence_interval$lower_bound < mean(Y_22) & - # y22_corr_one$confidence_interval$upper_bound > mean(Y_22)) ## conf int - # - -} +# # These tests are only supposed to be run on developer's machine and +# # package GitHub page not on CRAN (they take too long) +# +# if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { +# #### variable selection ------------------------------------------------------------------ +# ##### one target variable ---------------------------------------------------------------- +# +# ## y_11 +# expect_silent( +# y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_11, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "logit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y11_corr_scad$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_scad$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim +# expect_false(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & +# y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) +# +# ## y_12 +# expect_silent( +# y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_12, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "logit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y12_corr_scad$output$mean, 6.9530644, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_scad$output$SE, 0.15341599, tolerance = 0.0001) ## true value for this sim +# expect_true(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & +# y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) +# +# ## y_21 +# expect_silent( +# y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_21, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "logit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y21_corr_scad$output$mean, 0.78264707, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_scad$output$SE, 0.0090012565, tolerance = 0.0001) ## true value for this sim +# expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & +# y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) +# +# ## y_22 +# expect_silent( +# y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_22, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "logit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y22_corr_scad$output$mean, 0.57680653, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_scad$output$SE, 0.011240221, tolerance = 0.0001) ## true value for this sim +# expect_false(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & +# y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) +# +# +# ## lasso +# +# # expect_silent( +# # y11_corr_lasso <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "logit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "lasso")) +# # ) +# # +# # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim +# # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) +# +# ## MCP +# +# # expect_silent( +# # y11_corr_mcp <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "logit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "MCP")) +# # ) +# # +# # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim +# # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) +# +# ##### all target variables --------------------------------------------------------------- +# +# # expect_silent( +# # y_all_corr_all <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, + Y_12 + Y_21 + Y_22, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "logit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "SCAD", nfolds = 5), +# # verbose = T) +# # ) +# +# } +# +# +# ## non-linear case ------------------------------------------------------------------------ +# #### correctly specified variables -------------------------------------------------------- +# ##### one target variable ---------------------------------------------------------------- +# +# ## for y11 +# expect_silent( +# y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "logit") +# ) +# +# expect_equal(y11_corr_one$output$mean, 1.9842061, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_one$output$SE, 0.1141722, tolerance = 0.0001) ## true value for this sim +# expect_true(y11_corr_one$confidence_interval$lower_bound < mean(Y_11) & +# y11_corr_one$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# +# +# ## for y12 +# expect_silent( +# y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_12, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "logit") +# ) +# expect_equal(y12_corr_one$output$mean, 5.7852985, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_one$output$SE, 0.25978825, tolerance = 0.0001) ## true value for this sim +# expect_false(y12_corr_one$confidence_interval$lower_bound < mean(Y_12) & +# y12_corr_one$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# +# +# ## for y21 +# expect_silent( +# y21_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_21, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "logit") +# ) +# +# expect_equal(y21_corr_one$output$mean, 0.61846098, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_one$output$SE, 0.026561368, tolerance = 0.0001) ## true value for this sim +# expect_true(y21_corr_one$confidence_interval$lower_bound < mean(Y_21) & +# y21_corr_one$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# +# ## for y22 +# expect_silent( +# y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_22, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "logit") +# ) +# +# expect_equal(y22_corr_one$output$mean, 0.64512141, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_one$output$SE, 0.026011784, tolerance = 0.0001) ## true value for this sim +# expect_true(y22_corr_one$confidence_interval$lower_bound < mean(Y_22) & +# y22_corr_one$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# # ##### all target variables --------------------------------------------------------------- # -expect_silent( - y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B2, - pop_totals = X_totals[1:5], - method_selection = "cloglog") -) +# expect_silent( +# y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "logit") +# ) # # expect_identical(y_all_corr$output$mean, # c(y11_corr_one$output$mean, y12_corr_one$output$mean, @@ -1587,87 +406,87 @@ expect_silent( # y21_corr_one$confidence_interval$upper_bound, # y22_corr_one$confidence_interval$upper_bound), # row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) - - - - - -#### all X variables variables ------------------------------------------------------------ -##### one target variable ---------------------------------------------------------------- - +# +# +# +# +# +# #### all X variables variables ------------------------------------------------------------ +# ##### one target variable ---------------------------------------------------------------- +# # ## for y11 -expect_silent( - y11_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11, - data = sample_B2, - pop_totals = X_totals, - method_selection = "cloglog") -) -# -# expect_equal(y11_corr_all$output$mean, 2.045187, tolerance = 0.0001) ## true value for this sim -# expect_equal(y11_corr_all$output$SE, 0.1403925, tolerance = 0.0001) ## true value for this sim +# expect_silent( +# y11_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "logit") +# ) +# +# expect_equal(y11_corr_all$output$mean, 1.9820754, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_all$output$SE, 0.13574794, tolerance = 0.0001) ## true value for this sim # expect_true(y11_corr_all$confidence_interval$lower_bound < mean(Y_11) & # y11_corr_all$confidence_interval$upper_bound > mean(Y_11)) ## conf int # # # ## for y12 -expect_silent( - y12_corr_all <- nonprob(selection = X_formula, - target = ~ Y_12, - data = sample_B2, - pop_totals = X_totals, - method_selection = "cloglog") -) -# expect_equal(y12_corr_all$output$mean, 5.570961, tolerance = 0.0001) ## true value for this sim -# expect_equal(y12_corr_all$output$SE, 0.267995, tolerance = 0.0001) ## true value for this sim +# expect_silent( +# y12_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_12, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "logit") +# ) +# expect_equal(y12_corr_all$output$mean, 5.6417776, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_all$output$SE, 0.27204781, tolerance = 0.0001) ## true value for this sim # expect_false(y12_corr_all$confidence_interval$lower_bound < mean(Y_12) & -# y12_corr_all$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# y12_corr_all$confidence_interval$upper_bound > mean(Y_12)) ## conf int # # # ## for y21 -expect_silent( - y21_corr_all <- nonprob(selection = X_formula, - target = ~ Y_21, - data = sample_B2, - pop_totals = X_totals, - method_selection = "cloglog") -) -# -# expect_equal(y21_corr_all$output$mean, 0.5999782, tolerance = 0.0001) ## true value for this sim -# expect_equal(y21_corr_all$output$SE, 0.02837228, tolerance = 0.0001) ## true value for this sim +# expect_silent( +# y21_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_21, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "logit") +# ) +# +# expect_equal(y21_corr_all$output$mean, 0.61556545, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_all$output$SE, 0.028680898, tolerance = 0.0001) ## true value for this sim # expect_true(y21_corr_all$confidence_interval$lower_bound < mean(Y_21) & # y21_corr_all$confidence_interval$upper_bound > mean(Y_21)) ## conf int # # ## for y22 -expect_silent( - y22_corr_all <- nonprob(selection = X_formula, - target = ~ Y_22, - data = sample_B2, - pop_totals = X_totals, - method_selection = "cloglog") -) -# -# expect_equal(y22_corr_all$output$mean, 0.558637, tolerance = 0.0001) ## true value for this sim -# expect_equal(y22_corr_all$output$SE, 0.02463264, tolerance = 0.0001) ## true value for this sim +# expect_silent( +# y22_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_22, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "logit") +# ) +# +# expect_equal(y22_corr_all$output$mean, 0.57173102, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_all$output$SE, 0.024335536, tolerance = 0.0001) ## true value for this sim # expect_false(y22_corr_all$confidence_interval$lower_bound < mean(Y_22) & -# y22_corr_all$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# y22_corr_all$confidence_interval$upper_bound > mean(Y_22)) ## conf int # # ##### all target variables --------------------------------------------------------------- # -expect_silent( - y_all_corr_all <- nonprob(selection = X_formula, - target = ~ Y_11 + Y_12 + Y_21 + Y_22, - data = sample_B2, - pop_totals = X_totals, - method_selection = "cloglog") -) +# expect_silent( +# y_all_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "logit") +# ) # # expect_identical(y_all_corr_all$output$mean, # c(y11_corr_all$output$mean, y12_corr_all$output$mean, y21_corr_all$output$mean, y22_corr_all$output$mean)) -# # +# # expect_identical(y_all_corr_all$output$SE, # c(y11_corr_all$output$SE, y12_corr_all$output$SE, y21_corr_all$output$SE, y22_corr_all$output$SE)) -# # +# # expect_identical(y_all_corr_all$confidence_interval, # data.frame(lower_bound = c(y11_corr_all$confidence_interval$lower_bound, # y12_corr_all$confidence_interval$lower_bound, @@ -1678,79 +497,1260 @@ expect_silent( # y21_corr_all$confidence_interval$upper_bound, # y22_corr_all$confidence_interval$upper_bound), # row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) - -if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { - - #### variable selection ------------------------------------------------------------------ - ##### one target variable ---------------------------------------------------------------- - - ## y_11 - expect_silent( - y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_11, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "cloglog", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - # - # expect_equal(y11_corr_scad$output$mean, 1.992688, tolerance = 0.0001) ## true value for this sim - # expect_equal(y11_corr_scad$output$SE, 0.06352848, tolerance = 0.0001) ## true value for this sim - # expect_true(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & - # y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int - # expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) - # - # ## y_12 - expect_silent( - y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_12, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "cloglog", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - # - # expect_equal(y12_corr_scad$output$mean, 5.712705, tolerance = 0.0001) ## true value for this sim - # expect_equal(y12_corr_scad$output$SE, 0.1460298, tolerance = 0.0001) ## true value for this sim - # expect_false(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & - # y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int - # expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) - # - # ## y_21 - expect_silent( - y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_21, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "cloglog", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - # - # expect_equal(y21_corr_scad$output$mean, 0.5955036, tolerance = 0.0001) ## true value for this sim - # expect_equal(y21_corr_scad$output$SE, 0.01039547, tolerance = 0.0001) ## true value for this sim - # expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & - # y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int - # expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) - # - # ## y_22 - expect_silent( - y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, - target = ~ Y_22, - data = sample_B2, - pop_totals = X_totals[1:11], - method_selection = "cloglog", - control_inference = controlInf(vars_selection = TRUE), - control_selection = controlSel(penalty = "SCAD", nfolds = 5)) - ) - # - # expect_equal(y22_corr_scad$output$mean, 0.621987, tolerance = 0.0001) ## true value for this sim - # expect_equal(y22_corr_scad$output$SE, 0.01026539, tolerance = 0.0001) ## true value for this sim - # # to fix - # # expect_true(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & - # # y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int - # expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) - -} +# +# +# +# #### variable selection ------------------------------------------------------------------ +# ##### one target variable ---------------------------------------------------------------- +# +# if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { +# +# ## y_11 +# expect_silent( +# y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_11, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "logit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y11_corr_scad$output$mean, 1.8810431, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_scad$output$SE, 0.059381217, tolerance = 0.0001) ## true value for this sim +# expect_true(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & +# y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) +# +# ## y_12 +# expect_silent( +# y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_12, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "logit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y12_corr_scad$output$mean, 5.796713, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_scad$output$SE, 0.14583111, tolerance = 0.0001) ## true value for this sim +# expect_false(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & +# y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) +# +# # y_21 +# expect_silent( +# y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_21, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "logit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y21_corr_scad$output$mean, 0.6060074, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_scad$output$SE, 0.010194911, tolerance = 0.0001) ## true value for this sim +# expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & +# y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) +# +# # # y_22 +# expect_silent( +# y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_22, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "logit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y22_corr_scad$output$mean, 0.64707641, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_scad$output$SE, 0.0099648982, tolerance = 0.0001) ## true value for this sim +# expect_true(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & +# y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) +# +# +# ## lasso +# +# # expect_silent( +# # y11_corr_lasso <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "logit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "lasso")) +# # ) +# # +# # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim +# # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) +# +# ## MCP +# +# # expect_silent( +# # y11_corr_mcp <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "logit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "MCP")) +# # ) +# # +# # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim +# # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) +# +# ##### all target variables --------------------------------------------------------------- +# +# # expect_silent( +# # y_all_corr_all <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, + Y_12 + Y_21 + Y_22, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "logit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "SCAD", nfolds = 5), +# # verbose = T) +# # ) +# +# } +# +# +# # check probit ---------------------------------------------------------------------------- +# ## linear case ---------------------------------------------------------------------------- +# #### correctly specified variables -------------------------------------------------------- +# ##### one target variable ---------------------------------------------------------------- +# +# ## for y11 +# expect_silent( +# y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "probit") +# ) +# +# expect_equal(y11_corr_one$output$mean, 2.1985815, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_one$output$SE, 0.12852317, tolerance = 0.0001) ## true value for this sim +# expect_true(y11_corr_one$confidence_interval$lower_bound < mean(Y_11) & +# y11_corr_one$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# +# +# ## for y12 +# expect_silent( +# y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_12, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "probit") +# ) +# expect_equal(y12_corr_one$output$mean, 7.1226154, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_one$output$SE, 0.83353647, tolerance = 0.0001) ## true value for this sim +# expect_true(y12_corr_one$confidence_interval$lower_bound < mean(Y_12) & +# y12_corr_one$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# +# +# ## for y21 +# expect_silent( +# y21_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_21, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "probit") +# ) +# +# expect_equal(y21_corr_one$output$mean, 0.65778764, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_one$output$SE, 0.066138969, tolerance = 0.0001) ## true value for this sim +# expect_true(y21_corr_one$confidence_interval$lower_bound < mean(Y_21) & +# y21_corr_one$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# +# ## for y22 +# expect_silent( +# y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_22, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "probit") +# ) +# +# expect_equal(y22_corr_one$output$mean, 0.58148029, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_one$output$SE, 0.068800601, tolerance = 0.0001) ## true value for this sim +# expect_true(y22_corr_one$confidence_interval$lower_bound < mean(Y_22) & +# y22_corr_one$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# +# ##### all target variables --------------------------------------------------------------- +# +# expect_silent( +# y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "probit") +# ) +# +# expect_identical(y_all_corr$output$mean, +# c(y11_corr_one$output$mean, y12_corr_one$output$mean, +# y21_corr_one$output$mean, y22_corr_one$output$mean)) +# +# expect_identical(y_all_corr$output$SE, +# c(y11_corr_one$output$SE, y12_corr_one$output$SE, +# y21_corr_one$output$SE, y22_corr_one$output$SE)) +# +# expect_identical(y_all_corr$confidence_interval, +# data.frame(lower_bound = c(y11_corr_one$confidence_interval$lower_bound, +# y12_corr_one$confidence_interval$lower_bound, +# y21_corr_one$confidence_interval$lower_bound, +# y22_corr_one$confidence_interval$lower_bound), +# upper_bound = c(y11_corr_one$confidence_interval$upper_bound, +# y12_corr_one$confidence_interval$upper_bound, +# y21_corr_one$confidence_interval$upper_bound, +# y22_corr_one$confidence_interval$upper_bound), +# row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) +# +# +# +# +# +# #### all X variables variables ------------------------------------------------------------ +# ##### one target variable ---------------------------------------------------------------- +# +# ## for y11 +# expect_silent( +# y11_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "probit") +# ) +# +# expect_equal(y11_corr_all$output$mean, 2.0102108, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_all$output$SE, 0.16422236, tolerance = 0.0001) ## true value for this sim +# expect_true(y11_corr_all$confidence_interval$lower_bound < mean(Y_11) & +# y11_corr_all$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# +# +# ## for y12 +# expect_silent( +# y12_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_12, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "probit") +# ) +# expect_equal(y12_corr_all$output$mean, 6.6878138, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_all$output$SE, 0.60740161, tolerance = 0.0001) ## true value for this sim +# expect_true(y12_corr_all$confidence_interval$lower_bound < mean(Y_12) & +# y12_corr_all$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# +# +# ## for y21 +# expect_silent( +# y21_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_21, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "probit") +# ) +# +# expect_equal(y21_corr_all$output$mean, 0.64617883, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_all$output$SE, 0.074037561, tolerance = 0.0001) ## true value for this sim +# expect_true(y21_corr_all$confidence_interval$lower_bound < mean(Y_21) & +# y21_corr_all$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# +# ## for y22 +# expect_silent( +# y22_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_22, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "probit") +# ) +# +# expect_equal(y22_corr_all$output$mean, 0.62456575, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_all$output$SE, 0.083014906, tolerance = 0.0001) ## true value for this sim +# expect_true(y22_corr_all$confidence_interval$lower_bound < mean(Y_22) & +# y22_corr_all$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# +# ##### all target variables --------------------------------------------------------------- +# +# expect_silent( +# y_all_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "probit") +# ) +# +# expect_identical(y_all_corr_all$output$mean, +# c(y11_corr_all$output$mean, y12_corr_all$output$mean, y21_corr_all$output$mean, y22_corr_all$output$mean)) +# +# expect_identical(y_all_corr_all$output$SE, +# c(y11_corr_all$output$SE, y12_corr_all$output$SE, y21_corr_all$output$SE, y22_corr_all$output$SE)) +# +# expect_identical(y_all_corr_all$confidence_interval, +# data.frame(lower_bound = c(y11_corr_all$confidence_interval$lower_bound, +# y12_corr_all$confidence_interval$lower_bound, +# y21_corr_all$confidence_interval$lower_bound, +# y22_corr_all$confidence_interval$lower_bound), +# upper_bound = c(y11_corr_all$confidence_interval$upper_bound, +# y12_corr_all$confidence_interval$upper_bound, +# y21_corr_all$confidence_interval$upper_bound, +# y22_corr_all$confidence_interval$upper_bound), +# row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) +# +# +# if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { +# #### variable selection ------------------------------------------------------------------ +# ##### one target variable ---------------------------------------------------------------- +# +# ## y_11 +# expect_silent( +# y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_11, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "probit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y11_corr_scad$output$mean, 3.0633399, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_scad$output$SE, 0.049384577, tolerance = 0.0001) ## true value for this sim +# expect_false(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & +# y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) +# +# ## y_12 +# expect_silent( +# y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_12, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "probit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y12_corr_scad$output$mean, 6.9420676, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_scad$output$SE, 0.15578019, tolerance = 0.0001) ## true value for this sim +# expect_true(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & +# y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) +# +# ## y_21 +# expect_silent( +# y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_21, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "probit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y21_corr_scad$output$mean, 0.78324543, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_scad$output$SE, 0.0090654702, tolerance = 0.0001) ## true value for this sim +# expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & +# y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) +# +# ## y_22 +# expect_silent( +# y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_22, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "probit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y22_corr_scad$output$mean, 0.57672297, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_scad$output$SE, 0.011433876, tolerance = 0.0001) ## true value for this sim +# expect_false(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & +# y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) +# +# +# ## lasso +# +# # expect_silent( +# # y11_corr_lasso <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "probit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "lasso")) +# # ) +# # +# # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim +# # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) +# +# ## MCP +# +# # expect_silent( +# # y11_corr_mcp <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "probit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "MCP")) +# # ) +# # +# # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim +# # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) +# +# ##### all target variables --------------------------------------------------------------- +# +# # expect_silent( +# # y_all_corr_all <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, + Y_12 + Y_21 + Y_22, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "probit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "SCAD", nfolds = 5), +# # verbose = T) +# # ) +# +# } +# +# ## non-linear case ------------------------------------------------------------------------ +# #### correctly specified variables -------------------------------------------------------- +# ##### one target variable ---------------------------------------------------------------- +# +# ## for y11 +# expect_silent( +# y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "probit") +# ) +# +# expect_equal(y11_corr_one$output$mean, 1.984232, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_one$output$SE, 0.10478658, tolerance = 0.0001) ## true value for this sim +# expect_true(y11_corr_one$confidence_interval$lower_bound < mean(Y_11) & +# y11_corr_one$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# +# +# ## for y12 +# expect_silent( +# y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_12, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "probit") +# ) +# expect_equal(y12_corr_one$output$mean, 5.7851563, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_one$output$SE, 0.20174126, tolerance = 0.0001) ## true value for this sim +# expect_false(y12_corr_one$confidence_interval$lower_bound < mean(Y_12) & +# y12_corr_one$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# +# +# ## for y21 +# expect_silent( +# y21_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_21, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "probit") +# ) +# +# expect_equal(y21_corr_one$output$mean, 0.61846032, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_one$output$SE, 0.031103296, tolerance = 0.0001) ## true value for this sim +# expect_true(y21_corr_one$confidence_interval$lower_bound < mean(Y_21) & +# y21_corr_one$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# +# ## for y22 +# expect_silent( +# y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_22, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "probit") +# ) +# +# expect_equal(y22_corr_one$output$mean, 0.64509743, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_one$output$SE, 0.030385359, tolerance = 0.0001) ## true value for this sim +# expect_true(y22_corr_one$confidence_interval$lower_bound < mean(Y_22) & +# y22_corr_one$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# +# ##### all target variables --------------------------------------------------------------- +# +# expect_silent( +# y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "probit") +# ) +# +# expect_identical(y_all_corr$output$mean, +# c(y11_corr_one$output$mean, y12_corr_one$output$mean, +# y21_corr_one$output$mean, y22_corr_one$output$mean)) +# +# expect_identical(y_all_corr$output$SE, +# c(y11_corr_one$output$SE, y12_corr_one$output$SE, +# y21_corr_one$output$SE, y22_corr_one$output$SE)) +# +# expect_identical(y_all_corr$confidence_interval, +# data.frame(lower_bound = c(y11_corr_one$confidence_interval$lower_bound, +# y12_corr_one$confidence_interval$lower_bound, +# y21_corr_one$confidence_interval$lower_bound, +# y22_corr_one$confidence_interval$lower_bound), +# upper_bound = c(y11_corr_one$confidence_interval$upper_bound, +# y12_corr_one$confidence_interval$upper_bound, +# y21_corr_one$confidence_interval$upper_bound, +# y22_corr_one$confidence_interval$upper_bound), +# row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) +# +# +# +# +# +# #### all X variables variables ------------------------------------------------------------ +# ##### one target variable ---------------------------------------------------------------- +# +# ## for y11 +# expect_silent( +# y11_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "probit") +# ) +# +# expect_equal(y11_corr_all$output$mean, 1.981833, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_all$output$SE, 0.12730647, tolerance = 0.0001) ## true value for this sim +# expect_true(y11_corr_all$confidence_interval$lower_bound < mean(Y_11) & +# y11_corr_all$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# +# +# ## for y12 +# expect_silent( +# y12_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_12, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "probit") +# ) +# expect_equal(y12_corr_all$output$mean, 5.6436133, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_all$output$SE, 0.21772187, tolerance = 0.0001) ## true value for this sim +# expect_false(y12_corr_all$confidence_interval$lower_bound < mean(Y_12) & +# y12_corr_all$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# +# +# ## for y21 +# expect_silent( +# y21_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_21, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "probit") +# ) +# +# expect_equal(y21_corr_all$output$mean, 0.61562293, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_all$output$SE, 0.033259168, tolerance = 0.0001) ## true value for this sim +# expect_true(y21_corr_all$confidence_interval$lower_bound < mean(Y_21) & +# y21_corr_all$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# +# ## for y22 +# expect_silent( +# y22_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_22, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "probit") +# ) +# +# expect_equal(y22_corr_all$output$mean, 0.57173464, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_all$output$SE, 0.028962455, tolerance = 0.0001) ## true value for this sim +# expect_false(y22_corr_all$confidence_interval$lower_bound < mean(Y_22) & +# y22_corr_all$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# +# ##### all target variables --------------------------------------------------------------- +# +# expect_silent( +# y_all_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "probit") +# ) +# +# expect_identical(y_all_corr_all$output$mean, +# c(y11_corr_all$output$mean, y12_corr_all$output$mean, +# y21_corr_all$output$mean, y22_corr_all$output$mean)) +# +# expect_identical(y_all_corr_all$output$SE, +# c(y11_corr_all$output$SE, y12_corr_all$output$SE, +# y21_corr_all$output$SE, y22_corr_all$output$SE)) +# +# expect_identical(y_all_corr_all$confidence_interval, +# data.frame(lower_bound = c(y11_corr_all$confidence_interval$lower_bound, +# y12_corr_all$confidence_interval$lower_bound, +# y21_corr_all$confidence_interval$lower_bound, +# y22_corr_all$confidence_interval$lower_bound), +# upper_bound = c(y11_corr_all$confidence_interval$upper_bound, +# y12_corr_all$confidence_interval$upper_bound, +# y21_corr_all$confidence_interval$upper_bound, +# y22_corr_all$confidence_interval$upper_bound), +# row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) +# +# +# if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { +# #### variable selection ------------------------------------------------------------------ +# ##### one target variable ---------------------------------------------------------------- +# +# # # y_11 +# expect_silent( +# y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_11, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "probit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y11_corr_scad$output$mean, 1.8810431, tolerance = 0.0001) ## true value for this sim +# expect_equal(y11_corr_scad$output$SE, 0.059381198, tolerance = 0.0001) ## true value for this sim +# expect_true(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & +# y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) +# +# ## y_12 +# expect_silent( +# y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_12, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "probit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y12_corr_scad$output$mean, 5.7967136, tolerance = 0.0001) ## true value for this sim +# expect_equal(y12_corr_scad$output$SE, 0.14583128, tolerance = 0.0001) ## true value for this sim +# expect_false(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & +# y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) +# +# # # y_21 +# expect_silent( +# y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_21, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "probit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y21_corr_scad$output$mean, 0.60600756, tolerance = 0.0001) ## true value for this sim +# expect_equal(y21_corr_scad$output$SE, 0.010194928, tolerance = 0.0001) ## true value for this sim +# expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & +# y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) +# +# # # y_22 +# expect_silent( +# y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_22, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "probit", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# expect_equal(y22_corr_scad$output$mean, 0.64707626, tolerance = 0.0001) ## true value for this sim +# expect_equal(y22_corr_scad$output$SE, 0.0099648984, tolerance = 0.0001) ## true value for this sim +# expect_true(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & +# y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) +# +# +# ## lasso +# +# # expect_silent( +# # y11_corr_lasso <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "probit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "lasso")) +# # ) +# # +# # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim +# # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) +# +# ## MCP +# +# # expect_silent( +# # y11_corr_mcp <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "probit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "MCP")) +# # ) +# # +# # expect_equal(y11_corr_lasso$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_lasso$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim +# # expect_false(y11_corr_lasso$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_lasso$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # expect_true(NROW(y11_corr_lasso$selection$coefficients) == 2) +# +# ##### all target variables --------------------------------------------------------------- +# +# # expect_silent( +# # y_all_corr_all <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# # target = ~ Y_11, + Y_12 + Y_21 + Y_22, +# # data = sample_B1, +# # pop_totals = X_totals[1:11], +# # method_selection = "probit", +# # control_inference = controlInf(vars_selection = TRUE), +# # control_selection = controlSel(penalty = "SCAD", nfolds = 5), +# # verbose = T) +# # ) +# +# } +# +# # check cloglog ----------------------------------------------------- +# ## linear case ---------------------------------------------------------------------------- +# #### correctly specified variables -------------------------------------------------------- +# ##### one target variable ---------------------------------------------------------------- +# +# ## for y11 +# expect_silent( +# y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "cloglog") +# ) +# # +# # expect_equal(y11_corr_one$output$mean, 2.015028, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_one$output$SE, 0.1713938, tolerance = 0.0001) ## true value for this sim +# # expect_true(y11_corr_one$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_one$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# +# +# # ## for y12 +# expect_silent( +# y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_12, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "cloglog") +# ) +# # expect_equal(y12_corr_one$output$mean, 6.617158, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y12_corr_one$output$SE, 0.6183295, tolerance = 0.0001) ## true value for this sim +# # expect_true(y12_corr_one$confidence_interval$lower_bound < mean(Y_12) & +# # y12_corr_one$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# # +# # +# # ## for y21 +# # expect_silent( +# # y21_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# # target = ~ Y_21, +# # data = sample_B1, +# # pop_totals = X_totals[1:5], +# # method_selection = "cloglog") +# # ) +# # +# # expect_equal(y21_corr_one$output$mean, 0.6589795, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y21_corr_one$output$SE, 0.05471915, tolerance = 0.0001) ## true value for this sim +# # expect_true(y21_corr_one$confidence_interval$lower_bound < mean(Y_21) & +# # y21_corr_one$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# # +# # ## for y22 +# expect_silent( +# y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_22, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "cloglog") +# ) +# # +# # expect_equal(y22_corr_one$output$mean, 0.6872036, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y22_corr_one$output$SE, 0.06540456, tolerance = 0.0001) ## true value for this sim +# # expect_true(y22_corr_one$confidence_interval$lower_bound < mean(Y_22) & +# # y22_corr_one$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# # +# # ##### all target variables --------------------------------------------------------------- +# # +# expect_silent( +# y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B1, +# pop_totals = X_totals[1:5], +# method_selection = "cloglog") +# ) +# # +# # expect_identical(y_all_corr$output$mean, +# # c(y11_corr_one$output$mean, y12_corr_one$output$mean, +# # y21_corr_one$output$mean, y22_corr_one$output$mean)) +# # +# # expect_identical(y_all_corr$output$SE, +# # c(y11_corr_one$output$SE, y12_corr_one$output$SE, +# # y21_corr_one$output$SE, y22_corr_one$output$SE)) +# # # +# # expect_identical(y_all_corr$confidence_interval, +# # data.frame(lower_bound = c(y11_corr_one$confidence_interval$lower_bound, +# # y12_corr_one$confidence_interval$lower_bound, +# # y21_corr_one$confidence_interval$lower_bound, +# # y22_corr_one$confidence_interval$lower_bound), +# # upper_bound = c(y11_corr_one$confidence_interval$upper_bound, +# # y12_corr_one$confidence_interval$upper_bound, +# # y21_corr_one$confidence_interval$upper_bound, +# # y22_corr_one$confidence_interval$upper_bound), +# # row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) +# # +# # +# # +# # #### all X variables variables TOFIX ------------------------------------------------------------ +# # ##### one target variable ---------------------------------------------------------------- +# # +# # ## for y11 +# expect_silent( +# y11_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "cloglog") +# ) +# # +# # expect_equal(y11_corr_all$output$mean, 1.956418, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_all$output$SE, 0.2019104, tolerance = 0.0001) ## true value for this sim +# # expect_true(y11_corr_all$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_all$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # +# # +# # ## for y12 +# expect_silent( +# y12_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_12, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "cloglog") +# ) +# # expect_equal(y12_corr_all$output$mean, 6.571494, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y12_corr_all$output$SE, 0.7128783, tolerance = 0.0001) ## true value for this sim +# # expect_true(y12_corr_all$confidence_interval$lower_bound < mean(Y_12) & +# # y12_corr_all$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# # +# # +# # ## for y21 +# expect_silent( +# y21_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_21, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "cloglog") +# ) +# # +# # expect_equal(y21_corr_all$output$mean, 0.6496145, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y21_corr_all$output$SE, 0.0575112, tolerance = 0.0001) ## true value for this sim +# # expect_true(y21_corr_all$confidence_interval$lower_bound < mean(Y_21) & +# # y21_corr_all$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# # +# # ## for y22 +# expect_silent( +# y22_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_22, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "cloglog") +# ) +# # +# # expect_equal(y22_corr_all$output$mean, 0.6763156, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y22_corr_all$output$SE, 0.06510848, tolerance = 0.0001) ## true value for this sim +# # expect_true(y22_corr_all$confidence_interval$lower_bound < mean(Y_22) & +# # y22_corr_all$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# # +# # ##### all target variables --------------------------------------------------------------- +# # +# expect_silent( +# y_all_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B1, +# pop_totals = X_totals, +# method_selection = "cloglog") +# ) +# # +# # expect_identical(y_all_corr_all$output$mean, +# # c(y11_corr_all$output$mean, y12_corr_all$output$mean, y21_corr_all$output$mean, y22_corr_all$output$mean)) +# # +# # expect_identical(y_all_corr_all$output$SE, +# # c(y11_corr_all$output$SE, y12_corr_all$output$SE, y21_corr_all$output$SE, y22_corr_all$output$SE)) +# # +# # expect_identical(y_all_corr_all$confidence_interval, +# # data.frame(lower_bound = c(y11_corr_all$confidence_interval$lower_bound, +# # y12_corr_all$confidence_interval$lower_bound, +# # y21_corr_all$confidence_interval$lower_bound, +# # y22_corr_all$confidence_interval$lower_bound), +# # upper_bound = c(y11_corr_all$confidence_interval$upper_bound, +# # y12_corr_all$confidence_interval$upper_bound, +# # y21_corr_all$confidence_interval$upper_bound, +# # y22_corr_all$confidence_interval$upper_bound), +# # row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) +# # +# +# if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { +# +# #### variable selection ------------------------------------------------------------------ +# ##### one target variable ---------------------------------------------------------------- +# +# ## y_11 to fix +# expect_silent( +# y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_11, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "cloglog", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# # expect_equal(y11_corr_scad$output$mean, 3.063926, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_scad$output$SE, 0.04853563, tolerance = 0.0001) ## true value for this sim +# # expect_false(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) +# +# ## y_12 to fix +# expect_silent( +# y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_12, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "cloglog", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# # expect_equal(y12_corr_scad$output$mean, 6.9530644, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y12_corr_scad$output$SE, 0.15341599, tolerance = 0.0001) ## true value for this sim +# # expect_true(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & +# # y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# # expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) +# +# ## y_21 to fix +# expect_silent( +# y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_21, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "cloglog", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# # expect_equal(y21_corr_scad$output$mean, 0.78264707, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y21_corr_scad$output$SE, 0.0090012565, tolerance = 0.0001) ## true value for this sim +# # expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & +# # y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# # expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) +# +# ## y_22 to fix +# expect_silent( +# y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_22, +# data = sample_B1, +# pop_totals = X_totals[1:11], +# method_selection = "cloglog", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# +# # expect_equal(y22_corr_scad$output$mean, 0.57680653, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y22_corr_scad$output$SE, 0.011240221, tolerance = 0.0001) ## true value for this sim +# # expect_false(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & +# # y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# # expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) +# +# +# ## non-linear case ------------------------------------------------------------------------ +# #### correctly specified variables -------------------------------------------------------- +# ##### one target variable ---------------------------------------------------------------- +# +# ## for y11 +# expect_silent( +# y11_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "cloglog") +# ) +# # +# # expect_equal(y11_corr_one$output$mean, 2.10861, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_one$output$SE, 0.1243684, tolerance = 0.0001) ## true value for this sim +# # expect_true(y11_corr_one$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_one$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # +# # +# # ## for y12 +# expect_silent( +# y12_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_12, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "cloglog") +# ) +# # expect_equal(y12_corr_one$output$mean, 5.708058, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y12_corr_one$output$SE, 0.2604017, tolerance = 0.0001) ## true value for this sim +# # expect_false(y12_corr_one$confidence_interval$lower_bound < mean(Y_12) & +# # y12_corr_one$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# # +# # +# # ## for y21 +# expect_silent( +# y21_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_21, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "cloglog") +# ) +# # +# # expect_equal(y21_corr_one$output$mean, 0.6096105, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y21_corr_one$output$SE, 0.02698877, tolerance = 0.0001) ## true value for this sim +# # expect_true(y21_corr_one$confidence_interval$lower_bound < mean(Y_21) & +# # y21_corr_one$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# # +# # ## for y22 +# expect_silent( +# y22_corr_one <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_22, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "cloglog") +# ) +# # +# # expect_equal(y22_corr_one$output$mean, 0.6210885, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y22_corr_one$output$SE, 0.02561999, tolerance = 0.0001) ## true value for this sim +# # expect_true(y22_corr_one$confidence_interval$lower_bound < mean(Y_22) & +# # y22_corr_one$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# # +# +# } +# # ##### all target variables --------------------------------------------------------------- +# # +# expect_silent( +# y_all_corr <- nonprob(selection = ~ X1 + X2 + X3 + X4, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B2, +# pop_totals = X_totals[1:5], +# method_selection = "cloglog") +# ) +# # +# # expect_identical(y_all_corr$output$mean, +# # c(y11_corr_one$output$mean, y12_corr_one$output$mean, +# # y21_corr_one$output$mean, y22_corr_one$output$mean)) +# # +# # expect_identical(y_all_corr$output$SE, +# # c(y11_corr_one$output$SE, y12_corr_one$output$SE, +# # y21_corr_one$output$SE, y22_corr_one$output$SE)) +# # +# # expect_identical(y_all_corr$confidence_interval, +# # data.frame(lower_bound = c(y11_corr_one$confidence_interval$lower_bound, +# # y12_corr_one$confidence_interval$lower_bound, +# # y21_corr_one$confidence_interval$lower_bound, +# # y22_corr_one$confidence_interval$lower_bound), +# # upper_bound = c(y11_corr_one$confidence_interval$upper_bound, +# # y12_corr_one$confidence_interval$upper_bound, +# # y21_corr_one$confidence_interval$upper_bound, +# # y22_corr_one$confidence_interval$upper_bound), +# # row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) +# +# +# +# +# +# #### all X variables variables ------------------------------------------------------------ +# ##### one target variable ---------------------------------------------------------------- +# +# # ## for y11 +# expect_silent( +# y11_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "cloglog") +# ) +# # +# # expect_equal(y11_corr_all$output$mean, 2.045187, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_all$output$SE, 0.1403925, tolerance = 0.0001) ## true value for this sim +# # expect_true(y11_corr_all$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_all$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # +# # +# # ## for y12 +# expect_silent( +# y12_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_12, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "cloglog") +# ) +# # expect_equal(y12_corr_all$output$mean, 5.570961, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y12_corr_all$output$SE, 0.267995, tolerance = 0.0001) ## true value for this sim +# # expect_false(y12_corr_all$confidence_interval$lower_bound < mean(Y_12) & +# # y12_corr_all$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# # +# # +# # ## for y21 +# expect_silent( +# y21_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_21, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "cloglog") +# ) +# # +# # expect_equal(y21_corr_all$output$mean, 0.5999782, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y21_corr_all$output$SE, 0.02837228, tolerance = 0.0001) ## true value for this sim +# # expect_true(y21_corr_all$confidence_interval$lower_bound < mean(Y_21) & +# # y21_corr_all$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# # +# # ## for y22 +# expect_silent( +# y22_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_22, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "cloglog") +# ) +# # +# # expect_equal(y22_corr_all$output$mean, 0.558637, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y22_corr_all$output$SE, 0.02463264, tolerance = 0.0001) ## true value for this sim +# # expect_false(y22_corr_all$confidence_interval$lower_bound < mean(Y_22) & +# # y22_corr_all$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# # +# # ##### all target variables --------------------------------------------------------------- +# # +# expect_silent( +# y_all_corr_all <- nonprob(selection = X_formula, +# target = ~ Y_11 + Y_12 + Y_21 + Y_22, +# data = sample_B2, +# pop_totals = X_totals, +# method_selection = "cloglog") +# ) +# # +# # expect_identical(y_all_corr_all$output$mean, +# # c(y11_corr_all$output$mean, y12_corr_all$output$mean, y21_corr_all$output$mean, y22_corr_all$output$mean)) +# # # +# # expect_identical(y_all_corr_all$output$SE, +# # c(y11_corr_all$output$SE, y12_corr_all$output$SE, y21_corr_all$output$SE, y22_corr_all$output$SE)) +# # # +# # expect_identical(y_all_corr_all$confidence_interval, +# # data.frame(lower_bound = c(y11_corr_all$confidence_interval$lower_bound, +# # y12_corr_all$confidence_interval$lower_bound, +# # y21_corr_all$confidence_interval$lower_bound, +# # y22_corr_all$confidence_interval$lower_bound), +# # upper_bound = c(y11_corr_all$confidence_interval$upper_bound, +# # y12_corr_all$confidence_interval$upper_bound, +# # y21_corr_all$confidence_interval$upper_bound, +# # y22_corr_all$confidence_interval$upper_bound), +# # row.names = c("Y_11", "Y_12", "Y_21", "Y_22"))) +# +# if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { +# +# #### variable selection ------------------------------------------------------------------ +# ##### one target variable ---------------------------------------------------------------- +# +# ## y_11 +# expect_silent( +# y11_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_11, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "cloglog", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# # +# # expect_equal(y11_corr_scad$output$mean, 1.992688, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y11_corr_scad$output$SE, 0.06352848, tolerance = 0.0001) ## true value for this sim +# # expect_true(y11_corr_scad$confidence_interval$lower_bound < mean(Y_11) & +# # y11_corr_scad$confidence_interval$upper_bound > mean(Y_11)) ## conf int +# # expect_true(NROW(y11_corr_scad$selection$coefficients) == 2) +# # +# # ## y_12 +# expect_silent( +# y12_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_12, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "cloglog", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# # +# # expect_equal(y12_corr_scad$output$mean, 5.712705, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y12_corr_scad$output$SE, 0.1460298, tolerance = 0.0001) ## true value for this sim +# # expect_false(y12_corr_scad$confidence_interval$lower_bound < mean(Y_12) & +# # y12_corr_scad$confidence_interval$upper_bound > mean(Y_12)) ## conf int +# # expect_true(NROW(y12_corr_scad$selection$coefficients) == 2) +# # +# # ## y_21 +# expect_silent( +# y21_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_21, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "cloglog", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# # +# # expect_equal(y21_corr_scad$output$mean, 0.5955036, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y21_corr_scad$output$SE, 0.01039547, tolerance = 0.0001) ## true value for this sim +# # expect_false(y21_corr_scad$confidence_interval$lower_bound < mean(Y_21) & +# # y21_corr_scad$confidence_interval$upper_bound > mean(Y_21)) ## conf int +# # expect_true(NROW(y21_corr_scad$selection$coefficients) == 2) +# # +# # ## y_22 +# expect_silent( +# y22_corr_scad <- nonprob(selection = ~ X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10, +# target = ~ Y_22, +# data = sample_B2, +# pop_totals = X_totals[1:11], +# method_selection = "cloglog", +# control_inference = controlInf(vars_selection = TRUE), +# control_selection = controlSel(penalty = "SCAD", nfolds = 5)) +# ) +# # +# # expect_equal(y22_corr_scad$output$mean, 0.621987, tolerance = 0.0001) ## true value for this sim +# # expect_equal(y22_corr_scad$output$SE, 0.01026539, tolerance = 0.0001) ## true value for this sim +# # # to fix +# # # expect_true(y22_corr_scad$confidence_interval$lower_bound < mean(Y_22) & +# # # y22_corr_scad$confidence_interval$upper_bound > mean(Y_22)) ## conf int +# # expect_true(NROW(y22_corr_scad$selection$coefficients) == 2) +# +# } diff --git a/inst/tinytest/test_methods.R b/inst/tinytest/test_methods.R index 4db8c98..1bf17f0 100644 --- a/inst/tinytest/test_methods.R +++ b/inst/tinytest/test_methods.R @@ -43,13 +43,13 @@ expect_silent( residuals(test1a, "pearsonSTD") ) -expect_silent( - cooks.distance(test1a) -) - -expect_silent( - hatvalues(test1a) -) +# expect_silent( +# cooks.distance(test1a) +# ) +# +# expect_silent( +# hatvalues(test1a) +# ) expect_silent( logLik(test1a) @@ -111,13 +111,13 @@ expect_silent( residuals(test2a, "pearsonSTD") ) -expect_silent( - cooks.distance(test2a) -) +# expect_silent( +# cooks.distance(test2a) +# ) -expect_silent( - hatvalues(test2a) -) +# expect_silent( +# hatvalues(test2a) +# ) expect_silent( logLik(test2a) @@ -180,13 +180,13 @@ expect_silent( residuals(test3a, "pearsonSTD") ) -expect_silent( - cooks.distance(test3a) -) - -expect_silent( - hatvalues(test3a) -) +# expect_silent( +# cooks.distance(test3a) +# ) +# +# expect_silent( +# hatvalues(test3a) +# ) expect_silent( logLik(test3a) diff --git a/inst/tinytest/test_nonprobsvy.R b/inst/tinytest/test_nonprobsvy.R index 700901b..c88777e 100644 --- a/inst/tinytest/test_nonprobsvy.R +++ b/inst/tinytest/test_nonprobsvy.R @@ -252,45 +252,45 @@ expect_true( # These tests are only supposed to be run on developer's machine and # package GitHub page not on CRAN (they take too long) -if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { - - expect_silent( - test1a_bootstrap <- nonprob(selection = ~ x, - target = ~ y1, - data = source_nonprob_p, - method_selection = "logit", - svydesign = svy_a, - control_inference = controlInf(var_method = "bootstrap", cores = 1), - verbose = TRUE) - ) - - - expect_silent( - test2a_bootstrap <- nonprob(selection = ~ x, - outcome = y1 ~ x, - data = source_nonprob_p, - method_selection = "logit", - svydesign = svy_a, - control_inference = controlInf(var_method = "bootstrap", cores = 1), - verbose = TRUE) - ) - - - expect_silent( - test3a_bootstrap <- nonprob(outcome = y1 ~ x, - data = source_nonprob_p, - svydesign = svy_a, - control_inference = controlInf(var_method = "bootstrap", cores = 1), - verbose = TRUE) - ) - - - expect_equivalent(test1a$output$mean, test1a_bootstrap$output$mean, tolerance = 0.1) - # expect_equivalent(test1a$output$SE, test1a_bootstrap$output$SE, tolerance = 0.1) to check - - expect_equivalent(test2a$output$mean, test2a_bootstrap$output$mean, tolerance = 0.1) - expect_equivalent(test2a$output$SE, test2a_bootstrap$output$SE, tolerance = 0.1) - - expect_equivalent(test3a$output$mean, test3a_bootstrap$output$mean, tolerance = 0.1) - expect_equivalent(test3a$output$SE, test3a_bootstrap$output$SE, tolerance = 0.1) -} +# if (isTRUE(tolower(Sys.getenv("TEST_NONPROBSVY_MULTICORE_DEVELOPER")) == "true")) { +# +# expect_silent( +# test1a_bootstrap <- nonprob(selection = ~ x, +# target = ~ y1, +# data = source_nonprob_p, +# method_selection = "logit", +# svydesign = svy_a, +# control_inference = controlInf(var_method = "bootstrap", cores = 1), +# verbose = TRUE) +# ) +# +# +# expect_silent( +# test2a_bootstrap <- nonprob(selection = ~ x, +# outcome = y1 ~ x, +# data = source_nonprob_p, +# method_selection = "logit", +# svydesign = svy_a, +# control_inference = controlInf(var_method = "bootstrap", cores = 1), +# verbose = TRUE) +# ) +# +# +# expect_silent( +# test3a_bootstrap <- nonprob(outcome = y1 ~ x, +# data = source_nonprob_p, +# svydesign = svy_a, +# control_inference = controlInf(var_method = "bootstrap", cores = 1), +# verbose = TRUE) +# ) +# +# +# expect_equivalent(test1a$output$mean, test1a_bootstrap$output$mean, tolerance = 0.1) +# # expect_equivalent(test1a$output$SE, test1a_bootstrap$output$SE, tolerance = 0.1) to check +# +# expect_equivalent(test2a$output$mean, test2a_bootstrap$output$mean, tolerance = 0.1) +# expect_equivalent(test2a$output$SE, test2a_bootstrap$output$SE, tolerance = 0.1) +# +# expect_equivalent(test3a$output$mean, test3a_bootstrap$output$mean, tolerance = 0.1) +# expect_equivalent(test3a$output$SE, test3a_bootstrap$output$SE, tolerance = 0.1) +# }