I am new to this thread, so any help would be great. I need to optimize a neural network in MATLAB using GA. My network has an input [2x98] and [1x98], I tried to ask MATLAB for help, but I still don’t know what to do :( Therefore, any help would be appreciated. Thanks in advance.
Edit: I think I did not say what needs to be optimized, as Dan said in the 1st answer. I think the most important is the number of hidden neurons. And maybe the number of hidden layers and training parameters, such as the number of eras or so. Sorry for not providing enough information, I still found out about it.
If this is homework, do everything you have been taught in class.
Otherwise, completely remove the MLP. Vector regression support ( http://www.csie.ntu.edu.tw/~cjlin/libsvm/ ) is much more reliable in training for a wide range of problems and almost never encounters a stuck -a-local-minima often gets into training MLP , which forces you to solve the problem of optimizing the network topography in order to find a network that will actually train.
well, you need to be more specific about what you are trying to optimize. Is this the size of the hidden layer? Do you have a hidden layer? Is it an optimization of parameters (learning speed, kernel parameters)?
, (# , # ...), , , , , GA "" . , "" .
GA "" . , . , , , . , GA "" .
( , ...), GA. , , . , GA . GA ( , ANN).
. , GA, , ANN...
MATLAB , ( GATool, ) Neural Network Toolbox. .
:
-, .
, .
Matlab , . ANN, GA.
ANN . this.
, . MATLAB, .
Google Scholar. 5-6 . , .