I am trying to train my neural network multiple times (100 times) with hidden neurons range (1 to 100), then output the plots(Elapsed time vs Hidden neurons, MSE vs H neurons)

Question

Adeb 2023-5-4

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1958224-i-am-trying-to-train-my-neural-network-multiple-times-100-times-with-hidden-neurons-range-1-to-10

编辑： Anjaneyulu Bairi 2023-8-25

% Solve an Input-Output Fitting problem with a Neural Network

% Script generated by Neural Fitting app

% Created 03-May-2023 15:14:53

%

% This script assumes these variables are defined:

%

% inputONETWOTHREE - input data.

% outputThreeThree - target data.

x = inputONETWOTHREE;

t = outputThreeThree;

% Choose a Training Function

% For a list of all training functions type: help nntrain

% 'trainlm' is usually fastest.

% 'trainbr' takes longer but may be better for challenging problems.

% 'trainscg' uses less memory. Suitable in low memory situations.

trainFcn = 'trainlm'; % Levenberg-Marquardt backpropagation.

elapsedTime = (hiddenLayerSize);

% Create a Fitting Network

net = fitnet(hiddenLayerSize,trainFcn);

%net.layers{1}.transferFcn = 'tansig'; % hidden layer 1

% Setup Division of Data for Training, Validation, Testing

net.divideParam.trainRatio = 80/100;

net.divideParam.valRatio = 0/100;

net.divideParam.testRatio = 20/100;

for i = 1:100;

hiddenLayerSize = i;

tic;

% Train the Network

[net,tr] = train(net,x,t);

elapsedTime(i) = toc;

% Test the Network

y = net(x);

e = gsubtract(t,y);

performance = perform(net,t,y)

% View the Network

view(net)

% Plots

% Uncomment these lines to enable various plots.

%figure, plotperform(tr)

%figure, plottrainstate(tr)

%figure, ploterrhist(e)

%figure, plotregression(t,y)

%figure, plotfit(net,x,t)

figure;

plot(hiddenLayerSize,elapsedTime,'-ob');

xlabel('Number of Hidden Neurons');

ylabel('Elapsed Time (s)');

title('Elapsed Time vs Number of Hidden Neurons');

title('MSE vs Number of Hidden Neurons');

end

##INPUT,OUTPUT FILES ARE ATTACHED

2 个评论
显示无隐藏无

Ranjeet 2023-5-12

It would be helpful if you can provide the issue you are facing while your training process. You have mentioned that you are training a network, but the question is not clear that you need help in.

Adeb 2023-5-16

Thanks Ranjeet, see the attached image... the plot of the output (elapsed time vs no of hidden neurons) should be in a single line ascending and not the way its showing (see attached image)

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Anjaneyulu Bairi 2023-8-25

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1958224-i-am-trying-to-train-my-neural-network-multiple-times-100-times-with-hidden-neurons-range-1-to-10#answer_1293472

编辑：Anjaneyulu Bairi 2023-8-25

在 MATLAB Online 中打开

I ran your code and got some errors, to resolve those errors I made few changes and I am mentioning those changes below.

Error 1: “x”, ”t” are imported as tables into my workspace, So I converted them to matrix and transpose them.

Error 2: “hiddenLayerSize is not defined”.

Change 1: converted “x”,”t” to matrix and transposed it.

Change 2: I assigned value 100 to “hiddenLayerSize” as you are storing the values for 100th iteration.

Then I got the below results by running the code.

Please find the total code below.

% Solve an Input-Output Fitting problem with a Neural Network
% Script generated by Neural Fitting app
% Created 03-May-2023 15:14:53
%
% This script assumes these variables are defined:
%
%   inputONETWOTHREE - input data.
%   outputThreeThree - target data.
x = inputONETWOTHREE;
t = outputThreeThree;
%
%Changes stars here
x=x{:,:} % converting into matrix  
x=transpose(x)  % transpose the matrix
t=t{:,:}  % converting into matrix
t=transpose(t) % transpose the matrix
hiddenLayerSize=100; % assigning value 100 to hiddenLayerSize
%Changes end here
%
% Choose a Training Function
% For a list of all training functions type: help nntrain
% 'trainlm' is usually fastest.
% 'trainbr' takes longer but may be better for challenging problems.
% 'trainscg' uses less memory. Suitable in low memory situations.
trainFcn = 'trainlm';  % Levenberg-Marquardt backpropagation.
elapsedTime = (hiddenLayerSize);
% Create a Fitting Network
net = fitnet(hiddenLayerSize,trainFcn);
%net.layers{1}.transferFcn = 'tansig';   % hidden layer 1
        
% Setup Division of Data for Training, Validation, Testing
net.divideParam.trainRatio = 80/100;
net.divideParam.valRatio = 0/100;
net.divideParam.testRatio = 20/100;
        
for i = 1:100;
    hiddenLayerSize = i;
    tic;
    % Train the Network
    [net,tr] = train(net,x,t);
    elapsedTime(i) = toc;
    
    % Test the Network
    y = net(x);
    e = gsubtract(t,y);
    performance = perform(net,t,y)
        
    % View the Network
    view(net)
        
    % Plots
    % Uncomment these lines to enable various plots.
    %figure, plotperform(tr)
    %figure, plottrainstate(tr)
    %figure, ploterrhist(e)
    %figure, plotregression(t,y)
    %figure, plotfit(net,x,t)
    figure;
    plot(hiddenLayerSize,elapsedTime,'-ob');
    xlabel('Number of Hidden Neurons');
    ylabel('Elapsed Time (s)');
    title('Elapsed Time vs Number of Hidden Neurons');
    title('MSE vs Number of Hidden Neurons');
end