Closing as a strict requirement of {fabletools} is that the models all contain the same response variable(s).
Multivariate forecasting is possible using vars(y1, y2) on the LHS, however I would expect across() to produce the same models for different response variables.
Closing as a strict requirement of
{fabletools}is that the models all contain the same response variable(s). Multivariate forecasting is possible usingvars(y1, y2)on the LHS, however I would expectacross()to produce the same models for different response variables.