[FAKE] GMM IC PR for comment

[bdpedigo]

Reference Issues/PRs

What does this implement/fix? Explain your changes.

Any other comments?

[bdpedigo]

suggest making this match LassoLarsIC a bit closer, eg "Such criteria are useful to select the value of the regularization parameter by making a trade-off between the goodness of fit and the complexity of the model." could basically replace "regularization parameter" with "gaussian mixture parameters"

[bdpedigo]

not necessarily initialized with k-means, right?

[bdpedigo]

perhaps worth explaining the options, mainly i dont know what random_from_data is from this description

[bdpedigo]

also, is kmeans ++ not the default? if not, why not? i think it is in sklearn if i remember correctly

[tingshanL]

yeah, not sure, apparently kmeans is the default in GaussianMixture

[bdpedigo]

lasso lars IC calls this "criterion_"

[bdpedigo]

in lassolarsIC, there is no "sub-object" with the best model; rather the whole class just operates as if it is that model. does that make sense? while i cant speak for them, my guess is this is closer to what they'd be expecting

[tingshanL]

I add the attributes like weights_, means_ from GaussianMixture into GaussianMixtureIC, but I found that I still need to save the best model (I call best_estimator_ in the newest version) in order to all predict. Did I understand you correctly?

[bdpedigo]

not a property of GaussianMixture, recommend not storing

[bdpedigo]

i usually make any methods that dont access self into functions

[bdpedigo]

min value could be "min_components"?

[bdpedigo]

could you explain this?

[bdpedigo]

this all seems fine but just a suggestion - https://numpy.org/doc/stable/reference/generated/numpy.argmin.html
docs imply that for ties, argmin gives the first. so in other words if results are sorted in order of complexity, just using argmin would do what you want. (can even leave a comment to this effect, if you go this route).

note that i think having the results sorted by complexity anyway is probably desireable?

[bdpedigo]

this is effectively a dictionary - recommend just using one, or a named tuple? i am just anti classes that only store data and dont have any methods, but that is just my style :)

[bdpedigo]

why not just use GridSearchCV as in their example? https://scikit-learn.org/stable/auto_examples/mixture/plot_gmm_selection.html#sphx-glr-auto-examples-mixture-plot-gmm-selection-py

it would abstract away some of the stuff you have to do to make parallel work, for instance

[github-actions]

❌ Linting issues

This PR is introducing linting issues. Here's a summary of the issues. Note that you can avoid having linting issues by enabling pre-commit hooks. Instructions to enable them can be found here.

You can see the details of the linting issues under the lint job here

---

ruff

ruff detected issues. Please run ruff --fix --output-format=full . locally, fix the remaining issues, and push the changes. Here you can see the detected issues. Note that the installed ruff version is ruff=0.5.1.

examples/linear_model/plot_tweedie_regression_insurance_claims.py:82:35: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
   |
81 |     # unquote string fields
82 |     for column_name in df.columns[df.dtypes.values == object]:
   |                                   ^^^^^^^^^^^^^^^^^^^^^^^^^^ E721
83 |         df[column_name] = df[column_name].str.strip("'")
84 |     return df.iloc[:n_samples]
   |

sklearn/cluster/_optics.py:327:12: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
    |
325 |         """
326 |         dtype = bool if self.metric in PAIRWISE_BOOLEAN_FUNCTIONS else float
327 |         if dtype == bool and X.dtype != bool:
    |            ^^^^^^^^^^^^^ E721
328 |             msg = (
329 |                 "Data will be converted to boolean for"
    |

sklearn/cluster/tests/test_dbscan.py:294:12: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
    |
292 |     obj = DBSCAN()
293 |     s = pickle.dumps(obj)
294 |     assert type(pickle.loads(s)) == obj.__class__
    |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E721
    |

sklearn/linear_model/tests/test_ridge.py:1023:12: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
     |
1022 |     assert len(ridge_cv.coef_.shape) == 1
1023 |     assert type(ridge_cv.intercept_) == np.float64
     |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E721
1024 | 
1025 |     cv = KFold(5)
     |

sklearn/linear_model/tests/test_ridge.py:1031:12: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
     |
1030 |     assert len(ridge_cv.coef_.shape) == 1
1031 |     assert type(ridge_cv.intercept_) == np.float64
     |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E721
     |

sklearn/metrics/pairwise.py:2391:12: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
     |
2389 |         dtype = bool if metric in PAIRWISE_BOOLEAN_FUNCTIONS else "infer_float"
2390 | 
2391 |         if dtype == bool and (X.dtype != bool or (Y is not None and Y.dtype != bool)):
     |            ^^^^^^^^^^^^^ E721
2392 |             msg = "Data was converted to boolean for metric %s" % metric
2393 |             warnings.warn(msg, DataConversionWarning)
     |

sklearn/model_selection/_split.py:2938:27: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
     |
2936 |                 if value is None and hasattr(self, "cvargs"):
2937 |                     value = self.cvargs.get(key, None)
2938 |             if len(w) and w[0].category == FutureWarning:
     |                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E721
2939 |                 # if the parameter is deprecated, don't show it
2940 |                 continue
     |

sklearn/model_selection/tests/test_validation.py:589:20: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
    |
588 |             # Make sure all the arrays are of np.ndarray type
589 |             assert type(cv_results["test_r2"]) == np.ndarray
    |                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E721
590 |             assert type(cv_results["test_neg_mean_squared_error"]) == np.ndarray
591 |             assert type(cv_results["fit_time"]) == np.ndarray
    |

sklearn/model_selection/tests/test_validation.py:590:20: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
    |
588 |             # Make sure all the arrays are of np.ndarray type
589 |             assert type(cv_results["test_r2"]) == np.ndarray
590 |             assert type(cv_results["test_neg_mean_squared_error"]) == np.ndarray
    |                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E721
591 |             assert type(cv_results["fit_time"]) == np.ndarray
592 |             assert type(cv_results["score_time"]) == np.ndarray
    |

sklearn/model_selection/tests/test_validation.py:591:20: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
    |
589 |             assert type(cv_results["test_r2"]) == np.ndarray
590 |             assert type(cv_results["test_neg_mean_squared_error"]) == np.ndarray
591 |             assert type(cv_results["fit_time"]) == np.ndarray
    |                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E721
592 |             assert type(cv_results["score_time"]) == np.ndarray
    |

sklearn/model_selection/tests/test_validation.py:592:20: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
    |
590 |             assert type(cv_results["test_neg_mean_squared_error"]) == np.ndarray
591 |             assert type(cv_results["fit_time"]) == np.ndarray
592 |             assert type(cv_results["score_time"]) == np.ndarray
    |                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E721
593 | 
594 |             # Ensure all the times are within sane limits
    |

sklearn/utils/estimator_checks.py:1504:8: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
     |
1503 |     # func can output tuple (e.g. score_samples)
1504 |     if type(result_full) == tuple:
     |        ^^^^^^^^^^^^^^^^^^^^^^^^^^ E721
1505 |         result_full = result_full[0]
1506 |         result_by_batch = list(map(lambda x: x[0], result_by_batch))
     |

sklearn/utils/tests/test_validation.py:1344:12: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
     |
1342 |         )
1343 |     assert str(raised_error.value) == str(err_msg)
1344 |     assert type(raised_error.value) == type(err_msg)
     |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E721
     |

sklearn/utils/validation.py:882:49: E721 Use `is` and `is not` for type comparisons, or `isinstance()` for isinstance checks
    |
880 |         if all(isinstance(dtype_iter, np.dtype) for dtype_iter in dtypes_orig):
881 |             dtype_orig = np.result_type(*dtypes_orig)
882 |         elif pandas_requires_conversion and any(d == object for d in dtypes_orig):
    |                                                 ^^^^^^^^^^^ E721
883 |             # Force object if any of the dtypes is an object
884 |             dtype_orig = object
    |

Found 14 errors.

_{Generated for commit: e2f9a77. Link to the linter CI: here}