Optimize a parameter in statistic forecast
We are trying the Statistic forecast.
In monthly model, we need to set parameters to run the statistic model.
(Like Holt's α in Holt's Winters model)
But this setup is manual input.
Can we calculate the best parameter(0.0~1.0) by Optimize and minimize MAPE?
We can optimize the results of alfa value in holts model
Copy the holts model for 9 time and fix the ALFA value from 0.1 to 0.9 , calculate MAPE for reach exercise and them pick the model giving the lowest MAPE , that would be most optimimum ALFA value to choose.
You can also setup an optimizer on Alfa , being the variable and build a minimization function for least MAPE