In order to find the most suitable parameters, I decided to use Solver (optimization tool in Excel).

However, you need to first add it to your Excel; explanation how to do it is here (2007) and here (2010)

Now the important part is explaining how the optimization actually works.

1) We want to minimize the Mean Absolute Percentage Error (MAPE) of this training set, so Set Target Cell is $HS1027

2) The Changing Cells are alpha, beta gamma cells so $O$2:$Q$2; so the solver is changing the values of those 3 parameters in order to find the lowest MAPE

3) However the parameters have to have Constraints:

a) 0<alpha<1 – but I used higher constraint 0,01<alpha<0,99

b) 0<beta<1 – but I used higher constraint 0,01<alpha<0,99

c) 0<gamma<1 – but I used higher constraint 0,01<alpha<0,99

4) Analogically I did with Multiplicative approach; the only thing that was changed was Target Cell (into $L1027$) and Changing Cells : $O$3:$Q$3

If you want to return to previous page, click: here