I've been looking for answers on this question but couldn't find concrete solutions so wanted to ask y'all.
I have been playing around trying to forecast an economic/financial-related indicator with a large number of variables, so I used Partial Least Squares instead of linear. However, this said indicator and some of the predictors are very seasonal. One thing I tried was to Loess decompose the seasonality and use the seasonal component ITSELF as a predictor, but i'm not sure if that's methodologically sound - is this allowed? What are other ways to factor in seasonality in PLS forecast?