Fix fetch_openml broken pandas dependency (#87)

* Add pam algorithm

* pam algorithm, not naive.

* black reformat

* Fix mistake in code

* optimization of the algorithm for speed, review from @kno10

* remove generator for couples

* fix mistake

* Update pam review 2

* fix mistake

* cython implementation

* add test

* disable openmp for windows and mac

* fix black

* fix setup.py for windows

* remove test

* change review

* fix black

* Add build, remove parallel computing

* Apply suggestions from code review

Co-authored-by: Roman Yurchak <[email protected]>

* apply suggested change & rename alternating to alternate.

* fix test

* make build default. Allow max_iter = 0 for build-only algo

* Test for method and init

* test on blobs example

* fix typo

* fix difference long/long long windows vs linux

* try another fix for windows/linux long difference

* test another fix cython long/int on different platforms

* test all in int, cython kmedoid

* Add doc cluster module, update doc kmedoids, change examples.

* fix black

* fix openml

* fix fetch_openml as_frame

Co-authored-by: Roman Yurchak <[email protected]>

Author: TimotheeMathieu

examples/robust/plot_robust_classification_diabete.py

@@ -21,7 +21,7 @@
 
 warnings.simplefilter(action="ignore", category=FutureWarning)
 
-X, y = fetch_openml(name="diabetes", return_X_y=True)
+X, y = fetch_openml(name="diabetes", as_frame=False, return_X_y=True)
 
 # replace the label names with 0 or 1
 y = (y == "tested_positive").astype(int)
@@ -83,3 +83,4 @@
 
 # Remark : when using accuracy score, the optimal hyperparameters change and
 #          for example the parameter c changes from 1.35 to 10.
+plt.show()

examples/robust/plot_robust_regression_california_houses.py

@@ -34,7 +34,7 @@ def quadratic_loss(est, X, y, X_test, y_test):
     return (est.predict(X_test) - y_test) ** 2
 
 
-X, y = fetch_california_housing(return_X_y=True)
+X, y = fetch_california_housing(as_frame=False, return_X_y=True)
 # Sub-sample for faster computation.
 X = X[:1000]
 y = y[:1000]