I have the following code while running i'm getting parse error at 'clear':usage might be invalid MATLAB syntax. I'm unable to fix this error could anyone help me..??

function nmse_calc = nmse(wbr, net, input, target)
% wbr contains the weights and biases vector in row vector form as 
% passed to it by the genetic algorithm. This must be transposed 
% when being set as the weights and biases vector for the network.
% net must be configured to the sizes of input and target
% input I x N matrix of N I-dimensional "I"nput vectors
% target O x N matrix of N O-dimensional "O"utput target vectors
%  vart1 = mean(var(target',1)); % Reference MSE
 net = setwb( net, wbr' );
output = net( input );
nmse_calc = mean( ( target(:) - output(:) ).^2 ) / vart1;
 end
clear all,
close all,
clc, 
plt=0;
tic
    [ x t ] = simplefit_dataset;
    [ I N ] = size(x);%[ 1 94 ]
    [O N ] = size(t);%[ 1 94 ]

% Reference MSE: Average Target Variance
    vart1 = mean(var(t',1)) ;% 8.3378 

% No. of Training Equations
    Ntrneq = N*O; % 94 

% No. of unknown weights for H hidden Nodes
% Nw = ( I +1 )*H + (H +1)*O

% No Overfitting Upper Bound for H
% Nw < Ntrneq <==> H < Hnoub
    Hnoub = ( Ntrneq - O ) / ( I + O + 1 ); % 31

% Simplefit Data Plot: Target vs Input
    plt = plt+1;
    figure( plt ), plot( x, t ), 

% No. of Smooth Plot Local Extrema
    NLE = 4;
% Hidden Node Bounds
% NLE <= H < Hnoub % 4 <= H <= 31
% Hidden Node Choice
    H = NLE;
    Nw = ( I + 1)*H + ( H +1)*O; % 4, 13

% Create Regression/Curve-Fitting Neural Network:
              net = fitnet( H );
             net.divideFcn = 'dividetrain'; % Nval = Ntst = 0

% Configure the Net for the Simplefit Dataset
               s = rng( 'default' );
               net = configure( net, x, t );

% Initial Weights and Normalized MSE 
               wbr = getwb( net )'; % row vector
               NMSE0 = mse( t - net( x ) ) / vart1;% 2.7851

% Create handle to the NMSE function, 
               hand = @(wbr) nmse( wbr, net, x, t );

% Set the Genetic Algorithm tolerance for minimum change in fitness function % before terminating algorithm to 1e-4 and display each iteration's results.

               gaopts = gaoptimset( 'TolFun', 1e-4, 'display', 'iter' );
              [ wbopt, errga ] = ga( hand, Nw, gaopts );

% Optimization terminated: average change in the fitness value less than 
% options.TolFun.
% Best Mean Stall
% Generation f-count f(x) f(x) Generations
% 146 29400 0.1345 1777 29

              errga = errga; % 0.13448
              wbopt = wbopt;

% wbopt = [13.532 -7.440 -24.08 -8.759 12.595 5.777 -21.553 
% 20.281 4.983 2.095 8.649 -1.585 -0.522 ]

        net = setwb( net, wbopt');
        NMSE = mse( t - net( x ) ) / vart1; % 0.13448
        totaltime = toc; % 690 sec (11.5 min)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% FOR COMPARISON 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

clear all, close all;
clc, plt = 0; tic
    [ x t ] = simplefit_dataset;
    [ I N ] = size(x); % [ 1 94 ]
    [O N ] = size(t); % [ 1 94 ]
% Reference MSE: Average Target Variance
    vart1 = mean(var(t',1)); % 8.3378 
% No. of Training Equations
    Ntrneq = N*O; % 94 
% No. of unknown weights for H hidden Nodes
% Nw = ( I +1 )*H + (H +1)*O
% No Overfitting Upper Bound for H
% Nw < Ntrneq <==> H < Hnoub
    Hnoub = ( Ntrneq - O ) /( I + O + 1 ); % 31
% Simplefit Data Plot: Target vs Input
    plt = plt+1;
    figure( plt ), plot( x, t ), 
% No. of Smooth Plot Local Extrema
    NLE = 4 ;
% Hidden Node Bounds
% NLE <= H < Hnoub % 4 <= H <= 31
% Hidden Node Choice
    H = NLE;
    Nw = ( I + 1)*H + ( H +1)*O; % 4, 13
% Create Regression/Curve-Fitting Neural Network:
    net = fitnet( H );
    net.divideFcn = 'dividetrain'; % Nval = Ntst = 0

% Configure the Net for the Simplefit Dataset

     s = rng( 'default' );
     net = configure( net, x, t ); 
     net = train(net,x,t);
     NMSE = mse( t - net( x ) ) / vart1; % 3.274e-4
     totaltime = toc; % 11 sec

[B.k Sumedha]

Specify the whole error.The part which is displayed by the red.

[Priya Dharshini]

Answer 1

[Greg Heath]

You did not copy the original code correctly:

http://www.mathworks.com/matlabcentral/newsreader/view_thread/344888#944621

1. The function nmse should be stored in a separate file

2. You divided the first line into separate lines ending in commas

Hope this helps.

Greg

Answer 2

[Greg Heath]

The Code for the function nmse (above the clear command) has to saved in a separate file.

[Priya Dharshini]

When i save the code for function nmse in a separate file i'm getting this error Not enough input arguments. Now should i rectify this error

[Stephen23]

@Priya Dharshini: the error message is quite clear: you are not providing enough input arguments to the function. The solution is also simple: provide all of the input arguments.

function nmse_calc = nmse(wbr, net, input, target)

There are four input arguments. How many acre you calling the function with?

Answer 3

[Walter Roberson]

You should not end a line with a comma unless the comma is commented out, or unless the comma occurs inside a "[" that has a matching "]" on the later line.

Wrong:

   clear all,

Acceptable:

   clear all %,

Better:

   clear all

Best:

   %we had a clear all here but we realized it was wrong and destructive

Which is to say, you should never have a "clear all" in your code. Having "clear all" in code is rather like Wile E. Coyote exploding the bridge underneath himself and expecting to stay safe in mid-air as long as he does not look down.

[Steven Lord]

Commas work just fine when separating commands if you want one of them to display something in the Command Window.

x = 1; y = 2; z = x+y; % None of these will display anything
x = 1, y = 2, z = x+y, % Each of these will display

USUALLY you're going to want to use semicolons to suppress display as in the first line of the example above, but occasionally you want to see the display.

[Walter Roberson]

However, comma at the end of a command is not defined.

Answer 4

[Steven Lord]

You have an extra END in your code immediately before the CLEAR call. That ends the main function in the file, so the CLEAR call (and everything after it) is outside the function. That is invalid, and when the MATLAB parser attempts to figure out what to do with those lines it can't. Either eliminate that END (if everything in that file is supposed to be part of one function) or move everything after that END into a separate file (if the later commands are not supposed to be part of that function) and your code should "work".

The reason I put "work" in quotes is because it's usually a Bad Idea to include "clear all" in a function. That will clear all the variables that exist at that time in the function workspace, which includes the input arguments with which you called the function! It's a little more acceptable in a script file, but I would still avoid it unless I was sure I absolutely needed to use it. I would probably eliminate the "clc" and "close all" calls as well. The latter two aren't as big a deal, but it's a little annoying when I run a function that I don't expect to clear the Command Window or close all my figures and the function does one or both of those operations.