this following back propagation algorithm is giving error

IP1 = input1; IP2 = input2; IP3 = input3; IP4 = input4; IP5 = input5; IP6 = input6; input1,input2-------input6 are of [6 6] size target = eye(36); P1=IP1; P2=IP2; P3=IP3; P4=IP4; P5=IP5; P6=IP6; T=target; %%PREMNMX function within the Neural Networks Toolbox to scale my data between 0 and 1 % r1 = premnmx(P1); % r2 = (r1+1)*(1-0)/2+0; %net = newff(minmax(P1),[16 36],{'logsig','logsig'},'traingdx','learngdm'); net = newff(minmax(P1), [16 36], {'logsig','logsig'} , 'trainlm' );

is giving error Error in ==> test1 at 20 net=train_nn(input1,input2,input3,input4,input5,input6,target);

[Image Analyst]

http://www.mathworks.com/matlabcentral/answers/13205-tutorial-how-to-format-your-question-with-markup

Answer 1

[Greg Heath]

newff has been obsolete since 2010. The minmax(input) version you are using is doubly obsolete.

What version of MATLAB and NNToolbox are you using?? use the command

 ver

There should be one input matrix (N columns of I dimensional input vectors) and one target matrix (N columns of O-dimensional output vectors).

There is no MATLAB function with name train_nn

 [ net tr output error ] = train(net,input,target);

Thank you for formally accepting my answer

Greg