I have start using PyCaret v3.0.x for Time Series Forecasting. I had pass on the data for a single store and single channel along with the transactions with data starting from 2021 with a frequency of month. The numbers seems to be way off as compared to the actuals. The forecasted value turn out to be 4662 as compared to the actual which is 3574 with a (1 - MAPE) of 69.65% which seems to be way off.
When i had run the compare_models() function, it returns the Decision Tree as the best model.
Any suggestions?
Regards
Adil
I am running the PyCaret v3.0.x for Time Series Forecasting, the expected transactions are way off as compared to the actuals after running the compare_model() functions.
