demonstrate SVM smoothing sin

data/math/svm.py

@@ -0,0 +1,30 @@
+#!/usr/bin/env python
+
+def sin():
+  """demonstrate SVM smoothing sin"""
+  import libsvm.svmutil as svm  # https://github.com/cjlin1/libsvm
+  import math
+  import matplotlib.pyplot as plt
+  import numpy as np
+
+  samples, fold = 100, 50
+  x = np.linspace (0, 6 * np.pi, samples)
+  features = list (map (lambda x: [x], x))  # [x1, x2, x3, ..] to [[x1], [x2], [x3], ..]
+  labels = list (map (lambda x: math.sin (x), x))
+  plt.subplot (2, 1, 1)
+  plt.plot (x, labels, label= 'sin')
+  plt.legend()
+  plt.grid (True)
+
+  prob = svm.svm_problem (labels[:fold], features[:fold])
+  m = svm.svm_train (prob, '-s 3 -t 2 -q')  # epsilon RBF
+  y, _acc, _vals = svm.svm_predict (labels, features, m, '-q')
+  plt.subplot (2, 1, 2)
+  plt.plot (x, y, label= 'svm')
+  plt.legend()
+  plt.grid (True)
+  plt.gcf().canvas.manager.set_window_title ('sin')
+  plt.tight_layout()
+  plt.show()
+
+sin()

data/math/xgboost-.py

@@ -99,7 +99,9 @@ def perfect_symmetry():
   ax.plot_trisurf(X[:, 0], X[:, 1], y_mul, cmap='viridis')
 
   dtrain = xgb.DMatrix(X, label=y_mul)
-  params = {'tree_method': 'gpu_hist'}
+  params = {
+    #'tree_method': 'gpu_hist'
+  }
   bst = xgb.train(params, dtrain, evals=[(dtrain, 'train')], num_boost_round=10)
   y_pred = bst.predict(dtrain)