dask PCA: use incremental PCA for better memory efficieny

markotoplak · Nov 6, 2023 · 215ef9b · 215ef9b
1 parent 17dc2e4
commit 215ef9b
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 4 deletions.
diff --git a/Orange/projection/pca.py b/Orange/projection/pca.py
@@ -285,8 +285,9 @@ def _initialize_wrapped(self, X=None, Y=None):
             else:
                 if params["iterated_power"] == "auto":
                     params["iterated_power"] = 0
+                params.pop("tol")
 
-                return dask_decomposition.PCA(**params)
+                return dask_decomposition.IncrementalPCA(**params)
 
         return self.__wraps__(**params)
 

diff --git a/Orange/tests/test_pca.py b/Orange/tests/test_pca.py
@@ -159,9 +159,8 @@ def test_improved_randomized_pca_sparse_data(self):
 
     @unittest.skipIf(sklearn_version.startswith('0.20'),
                      "https://github.com/scikit-learn/scikit-learn/issues/12234")
-    @with_dasktable
-    def test_incremental_pca(self, prepare_table):
-        data = prepare_table(self.ionosphere)
+    def test_incremental_pca(self):
+        data = self.ionosphere
         self.__ipca_test_helper(data, n_com=3, min_xpl_var=0.49)
         self.__ipca_test_helper(data, n_com=32, min_xpl_var=1)