MATLAB: Dividing the data into training,testing,validation

rbf

I have a dataset of 75×6,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

Best Answer

>> 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

Related Solutions

MATLAB: Neural network (fitnet) and data decomposition

The training and test indices are given in tr. Type, without semicolon

 tr = tr

to see what info is in tr.

Hope this helps.

Thank you for formally accepting my answer

Greg

MATLAB: Neural net toolbox: divideFcn = ”

In general you can specify

    1. The type of data division
    2. The trn/val/test ratios

If you don't specify anything, you will get the default of random data division with trn/val/test ratios 0.7/0.15/0.15.

If you specify ' ' or 'dividetrain', you will get no division i.e., 100/0/0.

If you only specify 'divideblock' the data will be divided into 3 solid blocks of trn/val/test with the default 0.7/0.15/0.15 ratios. However, you can also specify another set of block ratios. This is the datadivision I typically recommend for timeseries prediction.

The disappointing part of this thread is that you should have figured this out by

 1. Reading the help and doc documentations for the 
different datadivision options. 
 2. Demonstrating the options on a small data example. 
For example,
 input = 1:10;  target = input.^2;