MATLAB: How does matlab splits the data set into 3 sets for nnet

Deep Learning Toolboxneural networkneural networks

Hello The default function for mat lab nn toolbox is diverand which can also be seen in the GUI of nnstart. It splits the data set based on what percent we define the training,validation and testing.I want to ask does everytime we give the percentage for the testing training and validation the inputs are chosen entirely randomly ?

Suppose I have the input as 1234×400 and out put is 40×400 that represnt 40 class.

So it can be seen from the input that every 10 columns belongs to one class out of 40 .So If I choose training as 80% and validation as 10 and testing as 10 in the GUI so does it mean after frist 8 columns of each class will be chosen for training? And again if choose the same partition(80 10 and 10) some other 8 will be chosen for training ? means it is going to b randomly choosen always again and again ?Please explain me this.

Best Answer

Every time you train() the data will be chosen randomly if you are using random division.

Related Solutions

MATLAB: How to train a feedfordward neural network with error weights

 Use the command window
 help mse
 doc mse

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;