Currently the autoencoder.data.evaluate.AutoencoderEvaluator class is being used to "evaluate" the autoencoder's reconstruction error on a dataset. But this is simply a precursor to "anomaly detection".
Solution
Need to start looking at refactoring the AutoencoderEvaluator into the AnomalyDetector class ...
Problem
Currently the
autoencoder.data.evaluate.AutoencoderEvaluator
class is being used to "evaluate" the autoencoder's reconstruction error on a dataset. But this is simply a precursor to "anomaly detection".Solution
Need to start looking at refactoring the
AutoencoderEvaluator
into theAnomalyDetector
class ...