Anomaly Detection Submodule

This submodule contains relevant functionality to perform anomaly detection on a test dataset. The data split, configuration to correctly preprocess and reconstruct the dataset and the trained model are saved during training. The components are reconstructed in the manager file (ecgan.manager.detect()). The anomaly manager class (ecgan.anomaly_detection.anomaly_manager) performs the evaluation on the test data set.

• Anomaly Assessment
• Anomaly Manager
• Embedders
• Reconstruction
• Detectors
  • Base Detector
  • Classification Detector
  • Detector Factory
  • Reconstruction Detector