Executing huge output using less input by neural network

Hi,

I usually use the Neural Network toolbox of the MATLAB version 2010b. I want to write a program to predict different responses by neural network.

I am having input data (1x7 matrices) which I have used in Finite Element Method analysis as input. Target data is 3x1200 matrices which I have got from Finite Element Method analysis result. Is it possible to solve this type of problem by neural network? If not possible, does anyone have any suggestion to solve this problem?

Highly welcome for any informed instruction.

Answer 1

[Walter Roberson]

I do not see any particular reason you could not use Neural Networks for this purpose. I have no idea how much training would be required for it to be effective, though.

It is possible that memory might become an issue if you are using a 32 bit version of MATLAB; if so then I would estimate that a 64 bit version should work.

[Alhaz Uddin]

Dear Walter,

Thank you very much for your kind reply. I am using 64 bit version of MATLAB. What type of function I can use for this program: network, training function, transfer function or anything else? Could you please tell me elaborately how can I proceed?