Here is my problem: I have input [x1,..,xt,n1,..,nt,1,2,...,t] where there is a missing timestep xi, and I use neighboring time series (found with KNN) n1,...,nt to add more features, as well as time embeddings at the end. I want to predict the missing timestep xi using this input with a simple DNN regressor. Is there any useful way to use the time embeddings? I want xi+1 and xi-1 to have the most influence on the prediction of xi. Would positional encoding work for example?

I am staying away from RNNs because I want a simple regressor and my time series is very short.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.