perform command in neural netwrok
1 view (last 30 days)
Show older comments
I cant understand how perform function works in neural network! at the end of this page in section 5: https://www.mathworks.com/help/nnet/gs/classify-patterns-with-a-neural-network.html?requestedDomain=www.mathworks.com the performance is calculated but without any explanation! what does it mean? is it a good network? In other examples the performance value is 19! which one is better?
1 Comment
Greg Heath
on 18 Jul 2017
Ahhh. An ally in dismay !
https://www.mathworks.com/matlabcentral/newsreader/view_thread/349050#955175
Searching the NEWSGROUP using the search word 'QUICKIE' find
Compare those results with perform.
Hope this helps.
Greg
Accepted Answer
Greg Heath
on 18 Jul 2017
help perform
doc perform
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
net = patternnet;
performFcn = net.performFcn
% performFcn = 'crossentropy' AHA!
help crossentropy
doc crossentropy
% Still isn't clear? ... Join the club. I used to know what crossentropy was. However, when I gave technical talks (I'm an engineer, not a physicist) no one else knew or cared. What they wanted was error rates and Rsquare.
My preferences are
NORMALIZED MEAN SQUARE ERROR NMSE = 1-Rsquare %[ 0,1]
and
PER CENT ERROR RATE % [ 0, 100]
See my QUICKIES post
Hope this helps.
Thank you for formally accepting my answer
Greg
PS If you are a glutten for punishment, see GOOGLE
greg MATLAB crossentropy ==> 39,500 hits (;>)
otherwise see
HITS
NEWSGROUP ANSWERS
crossentropy 28 62
crossentropy greg 6 46
Hope this helps.
Thank you for formally accepting my answer
Greg
0 Comments
More Answers (0)
See Also
Categories
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!