MATLAB: Taraining error when neural network training is done .each time training isd one the results are different

MATLABneural netwrktraining error

i am doing neural network ,training the netwrk eah time gives different results.

Best Answer

Training performance varies because the default train/val/test data division AND initial weights are pseudorandom.

One of many solutions for sufficiently large data sets.

1. Initialize the RNG so that the same stream of pseudorandom numbers can be repeated.

rng(0)

2. Design 10 or more nets

3. Choose the net with the smallest validation (NOT TRAINING) set error.

4. Estimate the performance on unseen data with the test set error.

5. If performance is unsatisfactory, try increasing the number of hidden nodes.

How large are X and T?

Hope this helps.

Thank you for formally accepting my answer

Greg

Related Solutions

MATLAB: How to fix constant iteration in neural networks

When you train your net again, the random number generator is in a different state. Therefore you will have a different trn/val/tst split AND a different set of initial weights. The training will stop according to one of several stopping rules including

 1. performance goal achieved
 2. maximum epochs reached
 3. minimum gradient achieved
 4. maximum mu reached
 5. validation stop (validation performance reaches a local maximum)
 [ net tr y e ] = train(net,x,t)  % e=t-y
 stopcriterion  = tr.stop

or, if you are training in a double for loop

 stopcriteria{i,j} = tr.stop

This is great because all are chosen to optimize your performance. That is why every time I try a new candidate for H=number of hidden nodes, I design at least Ntrials = 10 nets. So, if I am considering 10 different values for H, I will have 100 designs which I summarize in 3 10 by10 matrices for training, validation and test performance.

The best net is determined from the nontraining validation set performance (smaller values of H are preferred) and an unbiased estimate of unseen nontraining data performance is obtained from the test set performance.

Hope this helps.

Thank you for officially choosing my answer

Greg

MATLAB: How do i know the specifications of that neural network?

How many classification categories, c ?

What is the dimensionality, I, of the input feature column vectors?

How many input column vectors for each class? N1=?,N2=?,...,Nc=?

Total number of input vectors N = sum(i=1:c){Ni}

Target vectors are N c-dimensional (0,1) unit column vectors from eye(c)

MATLAB classification function documentation:

 help patternnet
 doc patternnet

MATLAB classification examples:

 help nndatasets
 doc  nndatasets

Choose Iris flower classification example:

 [ x, t] = iris_dataset;
 [ I N ] = size( x)
 [ c N ] = size( t )
 truclassindices = vec2ind( t );

Test with the default No. of hidden nodes

 H = 10
 net = patternnet( H );
 rng('default')
 [ net tr y e ] = train( net, x, t );
 % y = net( x );
 % e  = t - y ;
 classindices = vec2ind( t );
 err          = classindices ~= truclassindices;
 Nerr         = sum(err)
 PctErr       = 100*Nerr/N
 tr           = tr    % No semicolon to print the training record to screen

% Individual train/val/test errors for each class can be readily obtained by extracting the individual train/val/test indices from the training record tr.

Hope this helps.

Thank you for formally accepting my answer

Greg

Best Answer

Related Solutions

MATLAB: How to fix constant iteration in neural networks

MATLAB: How do i know the specifications of that neural network?

Related Question