MATLAB: How to form an empirical equation using a trained Feed Forward Back Propagation Neural Network with Logsig transfer function, 2 input, 10 hidden, and 3 output neurons? Please anybody show me the procedure.

MATLABnntool

I'm doing a project based on modeling experimental data using Artificial Neural Networks. My Feed forward Back Propagation Network is based on the following architecture: 2 Input, 10 Hidden (Single layer), 3 Output Neurons with Logsig transfer function. I'm using nntool for the same. I've trained the network and obtained results with good accuracy but I'm confused on how to form an empirical equation for the above three outputs based on network weights. Please someone explain me the procedure.

Best Answer

Search BOTH the NEWSGROUP and ANSWERS using

 greg analytic

You should be able to find more than a few examples

Hope this helps.

Thank you for formally accepting my answer

Greg

Related Solutions

MATLAB: Elman Network with Logsig Output Neuron

net=newelm(elman_input,elman_targets,[ 5,5], {'tansig','tansig','logsig'});

The output dimension is automatically obtained from elman_targets.

MATLAB: Hey I need to know how to know how many neurons

Search the NEWSGROUP and ANSWERS using

 Ntrneq Nw

where I have used the criterion that the number of UNKNOWN weights, Nw, should not exceed the number of training EQUATIONS Ntrneq.

               No. of SEARCH HITS
            NEWSGROUP      ANSWERS
 Ntrneq Nw    43            152

This yields an upperbound, Hub, on the number of hidden nodes in a single hidden layer net.

If more hidden nodes are required, one can replace the default training function TRAINLM with TRAINBR

 help trainbr
 doc trainbr