How to use trainpso to train a neural network model

function [net, tr] = trainpso(net, tr, trainV, valV, testV, varargin)

the trainpso is shown like this, how to input the data so that i can use trainpso to train a neural network model?

Answer 1

[Darshan Ramakant Bhat]

The first argument for the function seems to be a neural network object. It will have a property called ' inputs' and ' outputs'. You can set your training inputs as

net.inputs=trainInput;
net.outputs=trainOutputs;

Also note that it expects the data in the form of cell arrays.

Another way is to directly configure your neural network with inputs and targets like below

net = configure(net,x,t);%x is input and t is target

Please refer the following link

http://www.mathworks.com/help/nnet/ref/train.html

Regards,

Darshan Bhat

[jiabo he]

trainpso is a training function like trainlm and trainscg. However, it is not in the training function list. There is some way online to combine it with trainlm to train ANN model, but I cannot figure it out. Thank you for your answer, anyway.

http://www.georgeevers.org/pso_research_toolbox.htm

The above toolbox and pso toolbox may be all I need to realize this function.