6 input and 12 output Neural Network

How to design 6 input and 12 output neural network?

I have an input dataset of 6x10000 and a target dataset of 12x10000.

I have tried using feedforwardnet, newft...with different combinations of hidden layers and neurons, have used tansig for hidden layers and purelin for output layer always.

Also tried using different training functions (trainc, trainoss, trainb, trains, traincgb etc)

I only get regression as 94% maximum in some cases.

What do I have to do for getting regression close to/or 99-100% ??

[Harsha M V]

My dataset is as below attached files..

[KSSV]

Try to remove the outliers and bad values. Did you try that?

[Harsha M V]

No I didnt, because all the data are generated by another software (LTSPICE).

[Greg Heath]

Therefore you will have to introduce a post generation stage that will deal with the outliers.

Hope this helps.

Greg

[Harsha M V]

Okay, I will try removing outliers.