v0.0.13

examples/run_S3VM.py

@@ -0,0 +1,37 @@
+from __future__ import absolute_import
+import numpy as np
+from sklearn import datasets
+from sklearn import metrics
+from sklearn.model_selection import train_test_split
+
+
+# normalization
+def normalize(x):
+	return (x - np.min(x))/(np.max(x) - np.min(x))
+
+def get_data():
+	X, y = datasets.load_breast_cancer(return_X_y=True)
+	X = normalize(X)
+	X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.6, random_state = 0)
+	rng = np.random.RandomState(42)
+	random_unlabeled_points = rng.rand(len(X_train)) < 0.1
+	y_train[random_unlabeled_points] = -1
+	# 
+	index, = np.where(y_train != -1)
+	label_X_train = X_train[index,:]
+	label_y_train = y_train[index]
+	index, = np.where(y_train == -1)
+	unlabel_X_train = X_train[index,:]
+	unlabel_y = -1*np.ones(unlabel_X_train.shape[0]).astype(int)
+	return label_X_train, label_y_train, unlabel_X_train, unlabel_y, X_test, y_test
+
+if __name__ == "__main__":
+	from semisupervised import S3VM
+
+	label_X_train, label_y_train, unlabel_X_train, unlabel_y, X_test, y_test = get_data()
+	# S3VM
+	model = S3VM()
+	model.fit(np.vstack((label_X_train,unlabel_X_train)), np.append(label_y_train, unlabel_y))
+	predict = model.predict(X_test)
+	acc = metrics.accuracy_score(y_test, predict)
+	print("S3VM accuracy", acc)

examples/run_pseudo_label_nn.py

@@ -8,6 +8,7 @@
 import numpy as np
 from sklearn import datasets
 from sklearn.model_selection import train_test_split
+from sklearn import metrics
 
 class DNN(object):
 	"""
@@ -64,9 +65,12 @@ def get_data():
 
 	from semisupervised.PseudoLabelSSL import PseudoCallback, PseudoLabelNeuralNetworkClassifier
 	pseudo_callback = PseudoCallback()
-	#print("pseudo_callback.pretrain", pseudo_callback.pretrain)
-	clf = PseudoLabelNeuralNetworkClassifier(DNNmodel, pseudo_callback)
-	clf.fit(np.vstack((label_X_train, unlabel_X_train)), np.append(label_y_train, unlabel_y))
+
+	model = PseudoLabelNeuralNetworkClassifier(DNNmodel, pseudo_callback)
+	model.fit(np.vstack((label_X_train, unlabel_X_train)), np.append(label_y_train, unlabel_y))
+	predict = model.predict(X_test)
+	acc = metrics.accuracy_score(y_test, predict)
+	print("pseudo-label accuracy", acc)