I wanna model a nonlinear dynamic system in which current and temperature are depend on each other. I want the output of the neural network recreate the current while being smart enough to know this output is temperature dependent as well.
How can I do it within Neural network toolbox of Matlab ?
It sounds like a nonlinear regression?? (http://hspm.sph.sc.edu/courses/J716/pdf/716-5%20Non-linear%20regression.pdf)
You may try the nonlinear system identification toolbox of Matlab:
m=nlarx(Data, Orders, sigmoidnet('num',10),'search','lm','MaxIter', 10);
Orders= [na nb nk]
sigmoidnet is a neural network.
http://www.mathworks.com/products/neural-network/code-examples.html
http://www.mathworks.com/help/nnet/examples.html?refresh=true
I think you want a temperature dependent distribution for output. If this is correct, it's better you Boltzmann machine not usual nn and you can't do this by matlab. But if temperature has specific dynamics, it's just two-dimensional data and you can use @Hector method.
who down-vote answer regularly in this topic?
- Badges
- Science topic