sanjay
Full Member level 1
learnlv2
Hello All,
I wasn't sure if I am posting under the right section regarding my problem/doubt with LVQ.
I am designing a LVQ network for classification problem, but on inspection of results that I obtain from classification, I observe that the LVQ network appears to be classyfing the patterns of my clusters by forming linear decision boundary. I am actually having my doubt if this is correct, because as per the documentation that I have come across, I have seen that it does has the ability to move the missclassified neuron away from the winning neuron. But so far I have failed to see this..
I have attached the zip files which contain the m files used by me in MATLAB.
clusters.m -> generates a set of clusters for classification
usage example [X,t,D] = clusters( 50, 1 ); X - input, t - targets
createLVQ.m -> design and train the LVQ
usage example [ netLVQ,tr ] = createLVQ( X', t );
I am trying to use 'learnlv2' algorithm as it provides the ability to move the missclassified neuron from the set of winning neuron's but just havent been able to obtain this. Can anyone please HELP ME or throw some suggestion/light on this topic
Thanking you for your time and effort
Hello All,
I wasn't sure if I am posting under the right section regarding my problem/doubt with LVQ.
I am designing a LVQ network for classification problem, but on inspection of results that I obtain from classification, I observe that the LVQ network appears to be classyfing the patterns of my clusters by forming linear decision boundary. I am actually having my doubt if this is correct, because as per the documentation that I have come across, I have seen that it does has the ability to move the missclassified neuron away from the winning neuron. But so far I have failed to see this..
I have attached the zip files which contain the m files used by me in MATLAB.
clusters.m -> generates a set of clusters for classification
usage example [X,t,D] = clusters( 50, 1 ); X - input, t - targets
createLVQ.m -> design and train the LVQ
usage example [ netLVQ,tr ] = createLVQ( X', t );
I am trying to use 'learnlv2' algorithm as it provides the ability to move the missclassified neuron from the set of winning neuron's but just havent been able to obtain this. Can anyone please HELP ME or throw some suggestion/light on this topic
Thanking you for your time and effort