MATLAB: [Time Series Tool] [ Learning Alghoritm ]

[neural time series tool] [training]MATLAB

Hi, I just saw that there are several learning alghoritm like:

* trainlm % Levenberg
* traingd % Polak 
* ..
* ..

In Neural Time series Tool Levenberg is the default.

Are there cases when I could use an another one?

My target is make prediction of power consume in a office, having some parameters like Temperature, irradiation, mounth ecc..

I just tried with traingdx to use momentum ad adaptive learning rate but I am obtaining bad results in confront of trainlm, but I dont know why.

I hope you can help me.

EDIT: I just tried that each alghoritm based on gradient descendent (traingd,a,traingdm) works very very bad and performance never improve through epoch! Why?

Enrico

Best Answer

Why don't you use all of the defaults unless you have a specific reason for not doing so.

Once you get a successful design then you can change a default and see the effect on your previous results.

Changing more than one default at a time may thwart your efforts at understanding.

Perhaps a good way to start is to use one of the MATLAB data sets that are used in the documentation.

Hope this helps.

Greg

Related Solutions

MATLAB: How to change the activation function in ANN model created using toolbox

You are approaching the problem in exactly the wrong way.

The multilayer perceptron with one hidden layer is a universal approximator. The only reason to use more than one hidden layer is to reduce the total number of unknown weights by reducing the total number of hidden nodes (i.e., H1+H2 < H).

Ntrn training pairs of I-dimensional inputs and O-dimensional output targets yields Ntrneq = Ntrn*O training equations. The best way to obtain a robust design that tends to be resistant to noise, interference, measurement and transcription error is to MINIMIZE the number of unknown weights that yield an acceptable solution. If possible, Nw << Ntrneq is desirable.

 1. Use FITNET (calls FEEDFORWARDNET) for regression and curve-fitting
 2. Use PATTERNNET (calls FEEDFORWARDNET) for classification and pattern-recognition
 3. You have a classification problem. Start with the simple code in
     help patternnet
     doc patternnet
 4. If there are c classes, the target matrix columns should be columns of eye(c): O = c.
 5. The relationship between trueclass indices 1:c and the target columns is
 target           = ind2vec(trueclassindices);
 trueclassindices = vec2ind(target);
 6. Before starting the design, get a "feel" for the data. This may include
  a. plot inputs
  b. plot targets
  c. plot targets vs inputs
  d. standardize inputs to zero mean and unit variance using zscore or mapstd.
  e. Repeat a and c 
  f. Remove or modify errors and outliers.
 7. Start simple with the example used in the help and doc documentation.
  help patternnet
  doc patternnet
 8. You only have to vary 2 things 
  a. Number of hidden nodes (want as small as feasible) 
  b. Initial random weights
 9. This can be accomplished with a double for loop as I have illustrated in zillions of examples in the NEWSGROUP and ANSWERS. Search results
 NEWSGROUP HITS
 greg patternnet Ntrials  8
 ANSWERS HITS
 greg patternnet Ntrials           60
 greg patternnet Ntrials  Hmax     22
 greg patternnet Ntrials  Hub      17
 greg patternnet Ntrials  Hub Hmax 10

Hope this helps.

Thank you for formally accepting my answer

Greg

MATLAB: [Matlab – Time series tool] How to read IE cross-correlation

If a net accurately models all of the salient deterministic characteristics of an input-output relationship, the error will not be significantly correlated with the input and the cross-correlation function will be indistinguishable from that of the input and random noise.

If a net accurately models all of the salient predictive characteristics of a time series lag relationship, the error will not be significantly autocorrelated and the auto-correlation function will be indistinguishable from that of random noise.

Thank you for formally accepting my answer.

Greg

Best Answer

Related Solutions

MATLAB: How to change the activation function in ANN model created using toolbox

MATLAB: [Matlab – Time series tool] How to read IE cross-correlation

Related Question