,the overallperformance in the confusion matrix changed although i used the same number of hidden layer

when i repeat the classification by nn and by the same number of hidden layer ,the overall-performance in the confusion matrix changed (96-100-98 ...etc) what is the wrong thanks and best regards

Answer 1

[Greg Heath]

Most likely you did not initialize the random number generator to the same initial state.

Hope this helps,

Thank you for formally accepting my answer

Greg

[yuyyy hygu]

load('PX.mat') load('T.mat') net = feedforwardnet(11); net = train(net,PX,T); view(net)

P3 = net (PX); plotconfusion(T,P3) this my code to classify tumor types can you help me how can intialize thanks

[Greg Heath]

Three of my favorites are

 rng(0)

 rng(default)

 rng(4151941)

For examples, search the NEWSGROUP and ANSWERS using

 greg fitnet rng

or

 greg patternnet rng

Hope this helps

Greg

[yuyyy hygu]

rng(default) i had been added this line only ,for my classifier then i got an acceptable results thanks

[yuyyy hygu]

Thanks Greg

[Greg Heath]

You are welcome.

Greg