Contradiction with Wine Classification Example in Artificial Neural Network Tool(2014)
조회 수: 2(최근 30일)
표시 이전 댓글
akde 2015년 5월 29일
답변: Greg Heath 2015년 5월 30일
I am following the example about the wine classification in the following link; http://nl.mathworks.com/help/nnet/examples/wine-classification.html
Instead of 'x' and 't' arrays I have 'train_set_input' and 'train_set_target' with dimensions 11x830 and 2x830 respectively
I create the net by using
net = patternnet(15);
then train it by
[net,tr] = train(net,train_set_input,train_set_target);
in the wine example the net is like the following;
which provides output as 1 or 2 or 3.
In my case I get the following net;
My outputs can be any value between 0 and 1.
What should I do to get discrete outputs as it is in the example? (instead of some value between 0 and 1, I want either 0 or 1).
PS(please note the differences between the output layers shown in the net. How can I get the similar net?)
Greg Heath 2015년 5월 30일
I don't know how you ended up with that configuration. However, it will work if you use vec2ind on the output to obtain the class indices.
Hope this helps.
Thank you for formally accepting my answer
Find more on Sequence and Numeric Feature Data Workflows in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!Start Hunting!