Merge pull request #5428 from VesnaT/dbscan_normalize

[ENH] DBSCAN: Optional normalization

Author: markotoplak

Orange/widgets/unsupervised/owdbscan.py

@@ -78,6 +78,7 @@ class Error(widget.OWWidget.Error):
     min_samples = Setting(4)
     eps = Setting(0.5)
     metric_idx = Setting(0)
+    normalize = Setting(True)
     auto_commit = Setting(True)
     k_distances = None
     cut_point = None
@@ -102,6 +103,8 @@ def __init__(self):
         gui.comboBox(box, self, "metric_idx",
                      items=list(zip(*self.METRICS))[0],
                      callback=self._metirc_changed)
+        gui.checkBox(box, self, "normalize", "Normalize features",
+                     callback=self._on_normalize_changed)
 
         gui.auto_apply(self.buttonsArea, self, "auto_commit")
         gui.rubber(self.controlArea)
@@ -161,9 +164,9 @@ def _compute_cut_point(self):
         self.cut_point = int(DEFAULT_CUT_POINT * len(self.k_distances))
         self.eps = self.k_distances[self.cut_point]
 
-        if self.eps < EPS_BOTTOM_LIMIT:
-            self.eps = np.min(
-                self.k_distances[self.k_distances >= EPS_BOTTOM_LIMIT])
+        mask = self.k_distances >= EPS_BOTTOM_LIMIT
+        if self.eps < EPS_BOTTOM_LIMIT and sum(mask):
+            self.eps = np.min(self.k_distances[mask])
             self.cut_point = self._find_nearest_dist(self.eps)
 
     @Inputs.data
@@ -180,13 +183,18 @@ def set_data(self, data):
         if self.data is None:
             return
 
-        # preprocess data
-        for pp in PREPROCESSORS:
-            self.data_normalized = pp(self.data_normalized)
+        self._preprocess_data()
 
         self._compute_and_plot()
         self.unconditional_commit()
 
+    def _preprocess_data(self):
+        self.data_normalized = self.data
+        for pp in PREPROCESSORS:
+            if isinstance(pp, Normalize) and not self.normalize:
+                continue
+            self.data_normalized = pp(self.data_normalized)
+
     def send_data(self):
         model = self.model
 
@@ -248,6 +256,13 @@ def _min_samples_changed(self):
         self._compute_and_plot(cut_point=self.cut_point)
         self._invalidate()
 
+    def _on_normalize_changed(self):
+        if not self.data:
+            return
+        self._preprocess_data()
+        self._compute_and_plot()
+        self._invalidate()
+
 
 if __name__ == "__main__":
     a = QApplication(sys.argv)

Orange/widgets/unsupervised/tests/test_owdbscan.py

@@ -1,9 +1,13 @@
 # pylint: disable=protected-access
+import unittest
+
 import numpy as np
 from scipy.sparse import csr_matrix, csc_matrix
 
 from Orange.data import Table
+from Orange.clustering import DBSCAN
 from Orange.distance import Euclidean
+from Orange.preprocess import Normalize, Continuize, SklImpute
 from Orange.widgets.tests.base import WidgetTest
 from Orange.widgets.tests.utils import simulate, possible_duplicate_table
 from Orange.widgets.unsupervised.owdbscan import OWDBSCAN, get_kth_distances
@@ -226,3 +230,45 @@ def test_missing_data(self):
         self.send_signal(w.Inputs.data, self.iris)
         output = self.get_output(w.Outputs.annotated_data)
         self.assertTupleEqual((150, 1), output[:, "Cluster"].metas.shape)
+
+    def test_normalize_data(self):
+        # not normalized
+        self.widget.controls.normalize.setChecked(False)
+
+        data = Table("heart_disease")
+        self.send_signal(self.widget.Inputs.data, data)
+
+        kwargs = {"eps": self.widget.eps,
+                  "min_samples": self.widget.min_samples,
+                  "metric": "euclidean"}
+        clusters = DBSCAN(**kwargs)(data)
+
+        output = self.get_output(self.widget.Outputs.annotated_data)
+        output_clusters = output.metas[:, 0]
+        output_clusters[np.isnan(output_clusters)] = -1
+        np.testing.assert_array_equal(output_clusters, clusters)
+
+        # normalized
+        self.widget.controls.normalize.setChecked(True)
+
+        kwargs = {"eps": self.widget.eps,
+                  "min_samples": self.widget.min_samples,
+                  "metric": "euclidean"}
+        for pp in (Continuize(), Normalize(), SklImpute()):
+            data = pp(data)
+        clusters = DBSCAN(**kwargs)(data)
+
+        output = self.get_output(self.widget.Outputs.annotated_data)
+        output_clusters = output.metas[:, 0]
+        output_clusters[np.isnan(output_clusters)] = -1
+        np.testing.assert_array_equal(output_clusters, clusters)
+
+    def test_normalize_changed(self):
+        self.send_signal(self.widget.Inputs.data, self.iris)
+        simulate.combobox_run_through_all(self.widget.controls.metric_idx)
+        self.widget.controls.normalize.setChecked(False)
+        simulate.combobox_run_through_all(self.widget.controls.metric_idx)
+
+
+if __name__ == '__main__':
+    unittest.main()

Orange/widgets/utils/slidergraph.py

@@ -104,6 +104,7 @@ def clear_plot(self):
         This function clears the plot and removes data.
         """
         self.clear()
+        self.setRange(xRange=(0.0, 1.0), yRange=(0.0, 1.0))
         self.plot_horlabel = []
         self.plot_horline = []
         self._line = None