Let's say that I have past data indicating how some time series panned out. Now I also have the beginnings of a new time series that I expect to pan out in a similar trend to the old one.

What are some general data science / machine learning / statistics techniques I can use to predict the rest of the new time series? What about using some sort of online model to iteratively predict the old model's trend, and then somehow applying that to the new model?

  • $\begingroup$ Your question isnt clear w.r.t what exactly do you mean new time series? It's best if you add an example of what you are talking about. $\endgroup$ Mar 15 '20 at 21:32

Here you go: https://otexts.com/fpp2

But my experience is you should start using GBM based models for time series forecasting to get really good forecasts.

Check out XGBoost, Prophet, BSTS models for time series forecasting.


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.