PLS: Add 'Inside Ellipsis' column to output data

Author: VesnaT

Orange/widgets/model/owpls.py

@@ -1,7 +1,10 @@
 from AnyQt.QtCore import Qt
+import numpy as np
+import scipy.stats as ss
 import scipy.sparse as sp
 
-from Orange.data import Table, Domain
+from Orange.data import Table, Domain, DiscreteVariable
+from Orange.data.util import get_unique_names
 from Orange.regression import PLSRegressionLearner
 from Orange.widgets import gui
 from Orange.widgets.settings import Setting
@@ -62,11 +65,32 @@ def update_model(self):
 
     def _create_output_data(self) -> Table:
         projection = self.model.project(self.data)
+        inside = np.full(len(self.data), np.nan)
+        if self.n_components > 1:
+            inside = self._inside_ellipsis(projection.X[:, :2])
         data_domain = self.data.domain
         proj_domain = projection.domain
         metas = proj_domain.metas + proj_domain.attributes
         domain = Domain(data_domain.attributes, data_domain.class_vars, metas)
-        return self.data.transform(domain)
+        data: Table = self.data.transform(domain)
+        inside_name = get_unique_names(domain, "Inside Ellipsis")
+        inside_var = DiscreteVariable(inside_name, values=("No", "Yes"))
+        return data.add_column(inside_var, inside, to_metas=True)
+
+    @staticmethod
+    def _inside_ellipsis(scores: np.ndarray) -> np.ndarray:
+        # https://stats.stackexchange.com/questions/577628/trying-to-understand-how-to-calculate-a-hotellings-t2-confidence-ellipse
+        # https://stackoverflow.com/questions/66179256/how-to-check-if-a-point-is-in-an-ellipse-in-python
+        assert scores.shape[1] > 1
+
+        f = ss.f.ppf(0.95, len(scores[0]), len(scores) - len(scores[0]))
+        m = [np.pi * x / 100 for x in range(201)]
+        cx = np.cos(m) * np.std(scores[:, 0]) * f
+        cy = np.sin(m) * np.std(scores[:, 1]) * f
+
+        a = ((scores[:, 0] - np.mean(cx)) ** 2) / (np.min(cx) ** 2)
+        b = ((np.mean(cy) - scores[:, 1]) ** 2) / (np.min(cy) ** 2)
+        return a + b <= 1
 
     @OWBaseLearner.Inputs.data
     def set_data(self, data):

Orange/widgets/model/tests/test_owpls.py

@@ -48,14 +48,14 @@ def test_output_data(self):
         output = self.get_output(self.widget.Outputs.data)
         self.assertEqual(output.X.shape, (506, 13))
         self.assertEqual(output.Y.shape, (506,))
-        self.assertEqual(output.metas.shape, (506, 5))
+        self.assertEqual(output.metas.shape, (506, 6))
 
     def test_output_data_multi_target(self):
         self.send_signal(self.widget.Inputs.data, self._data_multi_target)
         output = self.get_output(self.widget.Outputs.data)
         self.assertEqual(output.X.shape, (506, 12))
         self.assertEqual(output.Y.shape, (506, 2))
-        self.assertEqual(output.metas.shape, (506, 5))
+        self.assertEqual(output.metas.shape, (506, 6))
 
     def test_output_components(self):
         self.send_signal(self.widget.Inputs.data, self._data)
@@ -71,6 +71,15 @@ def test_output_components_multi_target(self):
         self.assertEqual(components.Y.shape, (2, 2))
         self.assertEqual(components.metas.shape, (2, 1))
 
+    def test_output_data_inside_ellipsis(self):
+        self.send_signal(self.widget.Inputs.data, self._data)
+        output: Table = self.get_output(self.widget.Outputs.data)
+        inside_col = list(output.get_column("Inside Ellipsis"))
+        inside = [1] * len(self._data)
+        for i in [162, 163, 257, 262, 364, 370, 374, 414]:
+            inside[i] = 0
+        self.assertEqual(inside, inside_col)
+
 
 if __name__ == "__main__":
     unittest.main()