I am using an LSTM Autoencoder model for time series anomaly detection. None of the anomalies get flagged because the reconstruction loss comes out to be zero for all data points on the clearly anomalous test dataset; however, the model does flag anomalies on the train dataset.

I have tried using better training data, callbacks to prevent overfitting, and STL to adjust the dataset for seasonal trends, but I still get zero loss and can't flag anomalies. I am considering using a window/time steps now.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.