MATLAB: Equation extraction in an Artificial Neural Network

annequation

I am a newbie in ANN's. I have a relatively large model which comprises of 15 independent variables (inputs) and one dependent variable (output). Since i do not know the exact relationship between the inputs and the output I am trying to use an ANN for curve fitting . I have a vast amount of training data so as to train the network. I would like to know how to extract the regression equation between the inputs and output because i want to know the exact mathematical relationship that describes my model. I have studied the answers, that have been in this forum, regarding the extraction of the regression equation but i want to know how to find the regression equation between the raw data (namely non normalized inputs, output) ?

Thank you

Best Answer

There is no way to use data to determine the analytic form of particular function without prior information.

For typical data, the analytic form for a universal approximator neural network is

 y = B2 + LW*tanh(B2+IW*x)

Hope this helps.

Greg

Related Solutions

MATLAB: Fit an equation with trained neural network

The single hidden layer feedforward net is a

     UNIVERSAL APPROXIMATOR!

In the default mode it's equation is

 y = B2 + LW*tanh(B1+IW*x)
 help fitnet
 doc fitnet

Hope this helps

Thank you for formally accepting my answer

Greg

MATLAB: Regressin equation from neural network does not match to net(x)

You did not take into account the default minmax normalization of inputs and targets

See

http://www.mathworks.com/matlabcentral/answers/80928-expression-of-ann-model

Hope this helps.

Greg

Best Answer

Related Solutions

MATLAB: Fit an equation with trained neural network

MATLAB: Regressin equation from neural network does not match to net(x)

Related Question