How to train RBF in neural network?

I want to know Radial Basis function in neural network is adaptive or not?Suppose i have some trained data and is it possible to add train data to it .Or need to train whole data ?

Answer 1

[Greg Heath]

Unfortunately, newrb does not allow you to specify initial configurations.

If your initial data set is huge. You may just want to add the old hidden node centers to the new data and train a new network.

Otherwise, combine the data sets and start anew.

Hope this helps.

Thank you for formally accepting my answer

Greg

P.S. Or, you could modify the source code to accept an initial configuration and send me a copy (;>)

[ebru temiz]

Hi Greg Health ; have got two questions hi i have been working about signature recocniation system in matlab.I decide to classify with rbfnn(radial basis function neural network) .I can do %60 in accuracy.The system must to work higher accuracy.Number of educational data is 728.(28*26).There are 28 persons signature.Each person has got 26 signatures.Number of test signature is 280.(28*10).There are 28 persons in signature recogniation system.Each person has got 10 signatures.

1-Which spread value i can chose? Code:

eg = 0.001; % sum-squared error goal sc = 11; % spread constant

newrb(X,T,GOAL,SPREAD,MN,DF) takes these arguments, X - RxQ matrix of Q input vectors. T - SxQ matrix of Q target class vectors. GOAL - Mean squared error goal, default = 0.0. SPREAD - Spread of radial basis functions, default = 1.0. MN - Maximum number of neurons, default is Q. DF - Number of neurons to add between displays, default = 25. and returns a new radial basis network.

The larger that SPREAD is the smoother the function approximation will be. Too large a spread means a lot of neurons will be required to fit a fast changing function. Too small a spread means many neurons will be required to fit a smooth function, and the network may not generalize well. Call newrb with different spreads to find the best value for a given problem. Here we design a radial basis network given inputs X and targets T. Code:

     X = [1 2 3];
     T = [2.0 4.1 5.9];
     net = newrb(X,T);
     Y = net(X)

my code is Code:

eg = 0.001; % sum-squared error goal sc = 11; % spread constant net = newrb(input,target,eg,sc);

There is'nt any parameters set nember of epochs How can i set number of epochs?

Please help me...

[Adnan]

Suppose you want 30 epochs then you have to write :

net = newrb(input,target,eg,sc,30);