function f=PredictValue(X,Y,Xyc)
[X_gy,minX, maxX,Y_gy,minY,maxY]=premnmx(X,Y);
k=2;
net=newff(minmax(X_gy),[k,1],
{'tansig' 'purelin'}, 'traincgf'); % net.performFcn
= 'sse';
net.performParam.ratio = 0.5;
net.trainParam.show = 5;
net.trainParam.epochs = 300;
net.trainParam.goal = 1e-5;
[net,tr]=train(net,X_gy,Y_gy);
Y_nh=sim(net,X_gy)
[Y_nh_hy]=postmnmx(Y_nh,minY,maxY);
Xyc_gy=tramnmx(Xyc,minX,maxX);
Xyc_yc_net=sim(net,Xyc);
[Xyc_yc_hy]=postmnmx(Xyc_yc_net,minY,maxY)
f=Xyc_yc_hy;
Next is the error Message:
MWMCR::EvaluateFunction error: Error using ==>
network.subsasgn at 489 "performFcn" cannot be set to non-
existing function "mse".
Any idea will be appreciated.
Thanks,
HL
hi,
I get the same error message. what did you do to solve the problem?
Thanks.