Function approximation but not with MLP. What can I use to get good results?

3 views (last 30 days)

Vecsei Gábor on 20 Mar 2016

Good result means: less than 18 MSE (mean squared error).

I've got an X matrix: 10000x9 (10000 rows, 9 dimensional input) I've got an Y matrix: 10000x1 (10000 rows, 1 dimensional output)

I tried to use MLP but I couldn't get less than 21 MSE.

So what else can I use for a function approximation problem?

Find the treasures in MATLAB Central and discover how the community can help you!