How to tune the regularization parameter in extreme learning machine?

I train ELM on my dataset and gives test accuracy of 92.13. I wonder how to tune parameters to gain more test accuracy?

[BERGHOUT Tarek]

if you want to get beter trainig accuracy in basic SLFN traind by ELM you should do this:

1- invrease the nuber of neurons one by one using incrimental ELM until the networks gives a stable accuracy

2- after that you can change the formula of beta "B=pinv(H)*T" to this one

(B=(inv((I/C)+H'*H)*H'*T);

wher I is the identity matrix you can type "help eye " in Matlab;

and try yo make your hidden layers as a squered matrix.

and after that you can increase your regularization parameter C incrimentaly until you get the minimam training error.

Answer 1

[BERGHOUT Tarek]

if you want to get beter trainig accuracy in basic SLFN traind by ELM you should do this:

1- invrease the nuber of neurons one by one using incrimental ELM until the networks gives a stable accuracy

2- after that you can change the formula of beta "B=pinv(H)*T" to this one

(B=(inv((I/C)+H'*H)*H'*T);

wher I is the identity matrix you can type "help eye " in Matlab;

and try yo make your hidden layers as a squered matrix.

and after that you can increase your regularization parameter C incrimentaly until you get the minimam training error.