0
$\begingroup$

I have built an LSTM autoencoder model to identify anomalies in time series wifi throughput data for over 100 customers. However, the definition of anomalies is very subjective. E.g. Customer A thinks that a wifi throughput of 10 mbps is an anomaly, whereas Customer B thinks 10 mbps is normal behavior. I have one model for all my customers. Can you tell me if there's a way using which I can customize anomalies for different customers? I was thinking if I could stack another model on top of the LSTM autoencoder that would take LSTM predictions as inputs and then customize them for different customers, i.e. whether or not to show a given anomaly to Customer A,B,C,...etc. The benefit of this approach would be that I would have to build, maintain and retrain only one model for all customers, but be able to customize the anomalies shown to different customers based on their feedback (thumbs up or thumbs down)

$\endgroup$

0

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.