Dividing the data into training,t​esting,val​idation

I have a dataset of 75x6,in which i want to divide the data into training ,testing and validation and use rbf neural network to classify them,please tell how to divide and classify using rbfneural network

i used newrbe for training and testing before ,but how to include validation data in it

for reference

http://imgur.com/h5kjH

please help

Answer 1

[Greg Heath]

>> lookfor divide

...

divideblock - Partition indices into three sets using blocks of indices.

divideind - Partition indices into three sets using specified indices.

divideint - Partition indices into three sets using interleaved indices.

dividerand - Partition indices into three sets using random indices.

dividetrain - Partition indices into training set only.

dividevec - Divide problem vectors into training, validation and test vectors.

>> help divideblock, doc divideblock ...

To use a function like newrbe with divided data:

1. Use the training design data to create several (10?) nets with different spread values.

2. Use the validation training set to choose the best net.

3. Return to 1 if you want to refine your search for an optimal spread value

4. Use the nondesign test set to predict performance on unseen nondesign data.

5. If the result is unsatifactory

 a. In order to reduce the bias of future test set predictions, 
obtain a new division of the data (perhaps with differet percentages).

 b. Return to step 1

Hope this helps

Thank you for accepting my answer.

Greg

[FIR]

Greg can u please help me with small set of coding. Can you please tell how to calculate training and validation accuracy and testing accuracy

[Greg Heath]

y = sim(net,x);

e = t-y;

MSE = mse(e)

NMSE = MSE/mean(var(t')) % approx

[FIR]

Sorry Greg for frustating you ,i tried but not getting the expected result how to perform ,I have 75x6 data,i want to perform rbf by dividing into training,validation and testing randomly,can u suggest extra code for this please

[Greg Heath]

Why didn't you post your code and the unexpected results?

[FIR]

for svm we have

svmStruct = svmtrain(xdata,group,'showplot',true);

for plotting graph

is there any function in rbf like this to plot graph

[Greg Heath]

I cannot help you if you

1. ignore my advice

2. take my advice, it doesn't work and you do not post the code and error messages

3. do not answer my questions.

Now, IN ADDITION to my previous questions:

Are you trying to use

newrbe (created but not trained ) or

newrb (simultaneously created and trained )

Neither uses validation and/or test sets.

As stated in my outline, they are used separately, after the creation command, using sim(net,xval) and sim(net,xtst)

To answer your last question you can see all of the net properties just

by entering the command (no semicolon)

net

Then look for plotFcn.

If a net doesn't exist your question can still be answered via

net = newrbe

net.plotFcn

Similarly for newrb.

[FIR]

my code is 

x=datats(:,1:2);

y=datats(:,3:4);

z=datats(:,5:6);

net = newrb(x,groups)

y = sim(net,x);

Error using ==> newrb at 86
Inputs and Targets have different numbers of columns.

Error in ==> examplerb at 32
net = newrb(x,groups)

[FIR]

I tried as per your code

net = newrb(x1,groups)

y = sim(net,xtst);

plotFcn

Undefined function or variable 'plotFcn'.

[Greg Heath]

Reread my instructions

Do not enter the command plotFcn.

Either

Enter the command net without the ending semicolon. Then look for plotFcn.

or

Enter the command

net.plotFcn

In fact, do both so that you will understand

[FIR]

Thanks i understood now