My understanding is that forecasts can become erratic when too many exogenous variables are introduced. Does anyone have data on what the upper limit of exogenous variables is?
c
Chris Gervais
09/23/2023, 1:00 PM
we're successfully training with thousands of exogenous variables (lagged and future) without issue
👀 1
if the model output is erratic, it's almost certainly your training + optimization process