Merge pull request #585 from RaymondY/development

Add PyTorch Version of VAE, AE, SO_GAAL based on the DL base detector && Fix some bugs in DL base detector.

examples/auto_encoder_example.py

@@ -1,7 +1,7 @@
 # -*- coding: utf-8 -*-
 """Example of using AutoEncoder for outlier detection
 """
-# Author: Yue Zhao <zhaoy@cmu.edu>
+# Author: Tiankai Yang <tiankaiy@usc.edu>
 # License: BSD 2 clause
 
 from __future__ import division
@@ -14,6 +14,7 @@
 # if pyod is installed, no need to use the following line
 sys.path.append(
     os.path.abspath(os.path.join(os.path.dirname("__file__"), '..')))
+sys.path.append(os.path.abspath(os.path.dirname("__file__")))
 
 from pyod.models.auto_encoder import AutoEncoder
 from pyod.utils.data import generate_data
@@ -35,7 +36,7 @@
 
     # train AutoEncoder detector
     clf_name = 'AutoEncoder'
-    clf = AutoEncoder(epochs=30, contamination=contamination)
+    clf = AutoEncoder(epoch_num=30, contamination=contamination)
     clf.fit(X_train)
 
     # get the prediction labels and outlier scores of the training data

examples/auto_encoder_torch_example.py → examples/so_gaal_example_new.py

@@ -1,8 +1,8 @@
 # -*- coding: utf-8 -*-
-"""Example of using Autoencoder for outlier detection (pytorch)
-detection
+"""Example of using Single-Objective Generative Adversarial Active
+Learning (SO_GAAL) for outlier detection
 """
-# Author: Yue Zhao <zhaoy@cmu.edu>
+# Author: Tiankai Yang <tiankaiy@usc.edu>
 # License: BSD 2 clause
 
 from __future__ import division
@@ -15,16 +15,16 @@
 # if pyod is installed, no need to use the following line
 sys.path.append(
     os.path.abspath(os.path.join(os.path.dirname("__file__"), '..')))
+sys.path.append(os.path.abspath(os.path.dirname("__file__")))
 
-from pyod.models.auto_encoder_torch import AutoEncoder
+from pyod.models.so_gaal_new import SO_GAAL
 from pyod.utils.data import generate_data
 from pyod.utils.data import evaluate_print
 
-
 if __name__ == "__main__":
     contamination = 0.1  # percentage of outliers
-    n_train = 20000  # number of training points
-    n_test = 2000  # number of testing points
+    n_train = 30000  # number of training points
+    n_test = 3000  # number of testing points
     n_features = 300  # number of features
 
     # Generate sample data
@@ -35,9 +35,9 @@
                       contamination=contamination,
                       random_state=42)
 
-    # train AutoEncoder detector
-    clf_name = 'AutoEncoder'
-    clf = AutoEncoder(epochs=10)
+    # train SO_GAAL detector
+    clf_name = 'SO_GAAL'
+    clf = SO_GAAL(epoch_num=6, contamination=contamination, verbose=2)
     clf.fit(X_train)
 
     # get the prediction labels and outlier scores of the training data

examples/vae_example.py

@@ -1,7 +1,7 @@
 # -*- coding: utf-8 -*-
 """Example of using Variational Auto Encoder for outlier detection
 """
-# Author: Andrij Vasylenko <[email protected]>
+# Author: Tiankai Yang <[email protected]>
 # License: BSD 2 clause
 
 from __future__ import division
@@ -14,6 +14,7 @@
 # if pyod is installed, no need to use the following line
 sys.path.append(
     os.path.abspath(os.path.join(os.path.dirname("__file__"), '..')))
+sys.path.append(os.path.abspath(os.path.dirname("__file__")))
 
 from pyod.models.vae import VAE
 from pyod.utils.data import generate_data
@@ -35,7 +36,7 @@
 
     # train VAE detector (Beta-VAE)
     clf_name = 'VAE'
-    clf = VAE(epochs=30, contamination=contamination, gamma=0.8, capacity=0.2)
+    clf = VAE(epoch_num=30, contamination=contamination, beta=0.8, capacity=0.2)
     clf.fit(X_train)
 
     # get the prediction labels and outlier scores of the training data