1

I am new to r programming language and using modeltime package for the first time. I am unable to find any documentation on this but wondering if anyone knows if it can forecast multivariate time series data? Could you point me to an example or documentation?

Thanks!

Rcurious
  • 11
  • 1
  • When it comes to dealing with multivariate datasets and the use of external regressors, there is no real modeltime documentation out there. As a statement below one of there documentations on their website they announce, that you will be able to do so, when you register to one of their courses. In the following link, you can see how a recipe for feature engineering is created: https://business-science.github.io/modeltime.ensemble/articles/getting-started-with-modeltime-ensemble.html – Leonhard Geisler Feb 08 '22 at 08:56
  • Thanks @LeonhardGeisler. I get the sense from the reference manual that one is able to incorporate external regressors but I cannot find an example. Their courses costs $700 :O – Rcurious Feb 08 '22 at 21:18
  • Also, @LeonhardGeisler - do you have a recommendation on a package that can do this in R? Thanks – Rcurious Feb 08 '22 at 21:22
  • When it comes to modeltime, maybe you need to use special xreg functions to prepare your recipe: https://business-science.github.io/modeltime/reference/recipe_helpers.html. Otherwise, you can take a look at the forecast and fable packages. There should be more public documentation available – Leonhard Geisler Feb 09 '22 at 12:29
  • in this post I present how it is simple to add external features into your recipes via "+": https://stackoverflow.com/questions/70655288/how-to-deal-with-external-regressors-in-time-series-recipes It is easier than I have thought! – Leonhard Geisler Feb 15 '22 at 13:28

0 Answers0