I tried to use the fbprophet model for time series forecasting of energy data. I tried to do cross validation of the prophet using certain initial, period and horizon parameters,then i did the performance metric , and found it to be that for a horizon of 7 days I get an mae of 200 , which is the lowest considering all the horizon values. How can i use this horizon value that gives me a much lower MAE compared to other horizon values in my final best prophet model? Is there a way to make use of the least score of the mae parameters, to make sure my final model has a lesser error score?….Thanks in advance.