How do I convert a Feed Forward Neural Network from version 2.0 to a later version?

I would like to convert a Feed Forward Neural Network from version 2.0 to a later version.

Answer 1

[MathWorks Support Team]

Use NNT2FF and TRAIN to update and train your network.

The NNT2FF function can be used to upgrade the network from version 2.0 to 3.0 or later versions. The syntax when using the NNT2FF function is as follows:

   net = nnt2ff(pr,{w1 w2 ...},{b1 b2 ...},{tf1 tf2 ...},btf,blr,pf)

NNT2FF(PR,{W1 W2 ...},{B1 B2 ...},{TF1 TF2 ...},BTF,BLR,PF) takes the following arguments and returns a feed-forward network:

        PR - Rx2 matrix of min and max values for R input elements.

        Wi  - Weight matrix for the ith layer.

        Bi  - Bias vector for the ith layer.

        TFi - Transfer function of ith layer, default = 'tansig'.

        BTF - Backprop network training function, default = 'traingdx'.

        BLF - Backprop weight/bias learning function, default = 'learngdm'.

        PF  - Performance function, default = 'mse'.

PLEASE NOTE: NNT2FF.m has one error.

Line 1 should read:

function net = nnt2ff(pr,w,b,tf,btf,blr,pf)

instead of:

function net = nnt2ff(pr,w,b,tf,btf,blr,bf)

Now, you can use the TRAIN function to train the network.