I have a time series dataset containing hourly data from a few year, like below. Let's assume that I want to make prediction for the next 3 hours (2021-01-01 19:00, 2021-01-01 20:00, 2021-01-01 21:00). The output from model (e.g. LSTM) is the following sequence: [Feature_1(t+1), Feature_1(t+2), Feature_1(t+3)]. One of the features of my training data is weather data.

Date Feature_1 Weather_feature
2019-01-01 15:00 value(t0) value(t0)
2019-01-01 16:00 value(t0) value(t0)
.... ... ...
2021-01-01 18:00 value(t0) value(t0)

Can I use data, that I already know in my model? I would like to use the weather forecast for the next hours to improve my predctions. For example at 2021-01-01 18:00 i know value Feature_1 and Weather_feature in time t(0) but also i know already Weather_feature in (t+1), (t+2), (t+3). How can i use it in the model?

At the moment, this is what one training sample looks like: [[Feature_1(t-n), Weather_feature(t-n)]...[Feature_1(t-1),Weather_feature(t-1), [Feature_1(t0), Weather_feature(t0)]]

  • $\begingroup$ Same problem here. Did you find a solution or some literature ? Thanks ! $\endgroup$
    – Theudbald
    Aug 22 at 12:36

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.