Change normalization to Scaling in SVM

Author: AndrejaKovacic

Orange/classification/svm.py

@@ -2,13 +2,13 @@
 
 from Orange.classification import SklLearner, SklModel
 from Orange.base import SklLearner as SklLearnerBase
-from Orange.preprocess import Normalize
+from Orange.preprocess import Scale
 
 __all__ = ["SVMLearner", "LinearSVMLearner", "NuSVMLearner",
            "OneClassSVMLearner"]
 
 
-svm_pps = SklLearner.preprocessors + [Normalize()]
+svm_pps = SklLearner.preprocessors + [Scale(center=Scale.NoCentering)]
 
 
 class SVMClassifier(SklModel):

Orange/widgets/model/owsvm.py

@@ -10,6 +10,8 @@
 from Orange.widgets.utils.owlearnerwidget import OWBaseLearner
 from Orange.widgets.utils.signals import Output
 from Orange.widgets.utils.widgetpreview import WidgetPreview
+from Orange.classification import SklLearner
+from Orange.preprocess import Scale
 
 
 class OWSVM(OWBaseLearner):
@@ -26,9 +28,13 @@ class OWSVM(OWBaseLearner):
 
     LEARNER = SVMLearner
 
+    pps = SklLearner.preprocessors
+    scaling = [Scale(center=Scale.NoCentering)]
+
     class Outputs(OWBaseLearner.Outputs):
         support_vectors = Output("Support vectors", Table, explicit=True)
 
+
     #: Different types of SVMs
     SVM, Nu_SVM = range(2)
     #: SVM type
@@ -56,6 +62,8 @@ class Outputs(OWBaseLearner.Outputs):
     limit_iter = Setting(True)
     #: maximum number of iterations
     max_iter = Setting(100)
+    #: scaling of data
+    scale_data = Setting(True)
 
     _default_gamma = "auto"
     kernels = (("Linear", "x⋅y"),
@@ -176,6 +184,10 @@ def _add_optimization_box(self):
             alignment=Qt.AlignRight, controlWidth=100,
             callback=self.settings_changed,
             checkCallback=self.settings_changed)
+        self.scaling_box = gui.checkBox(
+            self.optimization_box, self,
+            'scale_data', label='Scale data',
+            callback=self.settings_changed)
 
     def _show_right_kernel(self):
         enabled = [[False, False, False],  # linear
@@ -210,7 +222,7 @@ def create_learner(self):
             'probability': True,
             'tol': self.tol,
             'max_iter': self.max_iter if self.limit_iter else -1,
-            'preprocessors': self.preprocessors
+            'preprocessors': self.pps + self.scaling if self.scale_data else self.pps
         }
         if self.svm_type == self.SVM:
             return SVMLearner(C=self.C, epsilon=self.epsilon, **common_args)