how to create dataset for classification and pattern recognition (neural network)

Dear all I have a table of extracted features from the image. I need to create a dataset to use it in classification and pattern recognition. I have many images and i will group them according to disease stage to 4 or 5 groups or classes. could you help me how to create the dataset (inputs and targets) and if there some advice to create a good classifier and have a good results. thank you

[Walter Roberson]

Replaces http://uk.mathworks.com/matlabcentral/answers/236901-how-to-create-dataset-for-classification-and-pattern-recognition-neural-network

Answer 1

[Greg Heath]

For c classes: Create two matrices from N pairs of I-dimensional input vectors and c-dimensional {0,1} unit vector targets. The 1 in the target vectors indicates the class index integer (1:c). The relationship between the target matrix and the class index row vector is given by

 target = ind2vec(classindices)

 classindices = vec2ind(target)

 Use patternnet or the obsolete (but still available) newpr.

 See the examples in the help and doc documentation, e.g., 

 help patternnet

 doc patternnet

 Also see some of my posts in the NEWSGROUP and ANSWERS: 

 greg patternnet

Hope this helps.

Thank you for formally accepting my answer

Greg

[hanem ellethy]

thanks for your answer but still not clear for me I will follow your posts in the NEWSGROUP and ANSWERS and try to understand more. If not I will ask again. Best Regards

[Eric T]

It is sparse so to get it in a format you can read:

full(ind2vec([1 1 2 2 3 3]))