Hey everyone! I’ve been using Nixtla’s time series forecasting tools for my research and projects, and they’ve been incredibly helpful. I really appreciate Nixtla for your hard work and this community for being so supportive! I have a question about the exogenous variables in mlforecast. Does mlforecast support historical exogenous variables, similar to the hist_exog_list in neuralforecast? I understand that we can provide future values of exogenous features to MLForecast.predict via the X_df argument. However, I’m unsure if it also supports historical covariates. Any insights on this would be greatly appreciated. Thanks!

Hey. We have an example on how to do it here

That's very creative! Thanks!