I use Matlab neural network toolbox and I need someone to help me to get the best result and optimization, and suggest me the best optimization algorithm.My experimental data are numerical and qualitative contain of three factors and one target and I should combine these. I was trying to use binary numbers instead of qualitative data.
To get a helpfull answer you need to be more specific.
Do you need someone to write the code for you? Do you need help with the configuration of GA? Do you need help to choose the correct neural network structure? Do you want to know what tools you could use for that? I could go on...
Anyway you could check this out if you need a quick but not free solution
http://www.biocompsystems.com/products/ngo/
In neuran network knowledge presentatin is very important. large data some reduntant are rquired to removed so as to present only significant data to neural network. It is highly dependent on the specific problem,
Hello, you can combine your neural network code with a NSGA II algorithm in matlab. You can download the GA code at the following link : http://www.mathworks.com/matlabcentral/fileexchange/10429-nsga-ii-a-multi-objective-optimization-algorithm
I used this code with ANN toolbox to optimize wall physical characteristics in energy building simulation. Minor modifications have to be done to make both codes working together.
A full methodology is explained in the master thesis of Laurent Magnier (2008): http://spectrum.library.concordia.ca/976218/1/MR63229.pdf
Instead of using NSGA II, you could use GenOpt + Matlab, since your optimization problem is for only one objective function. You can download GenOpt at this link: http://simulationresearch.lbl.gov/GO/
If you have any questions, don't hesitate to contact me.
Mixing qualitative and quantitative outputs usually deteriorates the function of neural network.
Training the network with quantitative outputs of wide range is also difficult
- Badges
- Science topic