add default and random samplers

IndicoDataSolutions · Dec 11, 2017 · 4558d14 · 4558d14
1 parent 7ddd7e0
commit 4558d14
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 5 deletions.
diff --git a/enso/config.py b/enso/config.py
@@ -13,7 +13,7 @@
 DATA = {
     'Classify/AirlineComplaints',
     'Classify/Disaster',
-    'Classify/Irony',
+    #'Classify/Irony',
     # 'Classify/IMDB.small',
     # 'Classify/Economy',
     # 'Classify/Emotion',
@@ -53,7 +53,7 @@
 TEST_SETUP = {
     "train_sizes": [100, 250, 500, 1000],
     "n_splits": 5,
-    "samplers": ['Random', 'RandomA']
+    "samplers": ['Default', 'Random']
 }
 
 # Visualizations to display

diff --git a/enso/sample/__init__.py b/enso/sample/__init__.py
@@ -1,3 +1,6 @@
+from numpy.random import choice
+
+
 class Sampler(object):
     """Base class for all samples."""
 
@@ -16,11 +19,11 @@ def class_for(cls, sampler_string):
         raise ValueError("Invalid sampler attempted ({})".format(sampler_string))
 
 
-class Random(Sampler):
+class Default(Sampler):
     def sample(self):
         return self.data_points[0:self.train_size]
 
 
-class RandomA(Sampler):
+class Random(Sampler):
     def sample(self):
-        return self.data_points[(len(self.data_points)-self.train_size):]
+        return choice(self.data_points, self.train_size, replace=False)