Neural Network Pattern Recognition Tool.
Afficher commentaires plus anciens
Hi Everyone, I am quite new to neural networks. I have a problem with pattern recognition using Neural Network Pattern Recognition Tool. As it is shown in demos I created 2 data sets in order to perform simple classification task. Firs one (6x90) contains 90 column vectors with 6 shape coefficient each. Second data set - targets (2x90), contains 90 column vectors where each ith column indicates which category the ith shape (it's either circle or triangle) belongs to. I trained the network with default settings and saved my results. And I'm stuck at this point. How to use trained network to, for example, classify one column vector with 6 shape coefficients to one of two classes? I found tried this http://azload.com/artificial-neural-networks-the-tutorial-with-matlab-5768.html but it seems to be inappropriate for linearly nonseparable problems. Any help would be appreciated:) Jacek
Réponse acceptée
Plus de réponses (1)
Jacek
le 23 Jan 2011
0 votes
4 commentaires
Ram
le 24 Jan 2011
:)
Greg Heath
le 10 Fév 2013
If you should use c > 2 clases/categories, use columns of the c-dimensional unit matrix to indicate classes in the target matrix. Use ind2vec and vec2ind to transform between target/output matrices and class indices.
Greg
Sohail Sarwar
le 7 Avr 2016
Dear Greg, I want to categorize the data set of 1500 learners into four classes. Is that possible with nprtool? Since one of your posts says we can get result for two classes. Can you please elaborate above suggestion coz it can make things to work for me. Regards Sohail
Greg Heath
le 8 Avr 2016
I am not that familiar with nprtool.
However, I have posted many explanations of classifier design in BOTH the NEWSGROUP and ANSWERS especially using codes newff, newpr and patternnet.
Use the help and doc commands to obtain minimal code.
help patternnet
doc patternnet
Then to find more detailed examples.
greg patternnet
Hope this helps.
Greg
Catégories
En savoir plus sur Sequence and Numeric Feature Data Workflows dans Centre d'aide et File Exchange
Produits
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!