CODEWITHSUNDEEP
matlabneural-network
Anu
Anu

Reputation: 119

Retrain neural network after Ntrials

I am trying to make multiple designs of neural network and from these trials i want the best performing neural network and then retrain my final network with the best performing design. I get the best performance in trial 4. But when i try to retrain the network with these values, iget different outputs. Please help me. My code is:

load inputdata
load targetdata
 x=input;
 t=target;
hiddenLayerSize = 30;
net = patternnet(hiddenLayerSize);
net.trainFcn='trainbr';
net.trainParam.epochs=100;
rng(0);     
 Ntrials=20;
 for i = 1:Ntrials
   s{i} = rng;
   net  = configure(net,x,t);
   [ net tr y e] = train(net,x,t);
      % Best mseval over all epochs of ith run
            netIW0{i} = net.IW
            netb0{i}  = net.b
            netLW0{i} = net.LW
         tstind = tr.testInd;
         ytst = y(:,tstind);
         ttst = t(:,tstind);
         %mseval(i) = mse(net,ttst,ytst)
         mseval(i)=tr.best_tperf;
         plt=plt+1,figure(plt)
         plotconfusion(ttst,ytst)
          title([ ' TEST SET CONFUSION MATRIX. TRIAL =  ', num2str(i)] )
         hold off
 end
 [ minmseval ibest ] = min(mseval);
 rng=s{ibest};            % For repeating the best design
  bestnet = configure(net,x,t);
 [ bestnet tr y e] = train(net,x,t);
         bestIW0 = bestnet.IW
         bestb0  = bestnet.b
         bestLW0 = bestnet.LW
         msetst=tr.best_tperf;
         tstind = tr.testInd;
         ytst = y(:,tstind);
         ttst = t(:,tstind);
         fWb=getwb(bestnet);
         %msetst= mse(bestnet,ttst,ytst)
         plt=plt+1,figure(plt)
         plotconfusion(ttst,ytst)
         title([ ' OVERALL TEST SET CONFUSION MATRIX=  '] )
         hold off
         save bestnet
         view(bestnet)

Upvotes: 0

Views: 740

Answers (1)

vbar
vbar

Reputation: 81

I do not know why You can not reproduce Your best network. But You can work around the problem by saving every of this 20 nets as separate objects and than just recall the best one and delete others. And this will be a faster way because You do not need to train again Your network.

Upvotes: 1

Related Questions