Array indices must be positive integers or logical values, Neural Network Performances

Hi, I am trying to improve the neural network according to this documentation

https://ch.mathworks.com/help/deeplearning/ug/improve-neural-network-generalization-and-avoid-overfitting.html

% %   Einleger_Binaer_Alle_inv_sortiert - input data.
% %   Auslenkung_Alle_inv - target data.
% 
x = Einleger_Binaer_Alle_inv_Sortiert;
t = Auslenkung_Alle_inv;

rng('default');

% 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';  % nicht Levenberg-Marquardt backpropagation, da schneller

% Create a Fitting Network
% hiddenLayerSize = 10;
% network = fitnet(hiddenLayerSize,trainFcn);

% For More hidden Layers und Neurons
n = 3;
s = 10;
layerSizes = ones(1, n)*s;  % Creating vector with layer sizes
network = fitnet(layerSizes, trainFcn);
 

% Choose Input and Output Pre/Post-Processing Functions
% For a list of all processing functions type: help nnprocess
network.input.processFcns = {'removeconstantrows','mapminmax'};
network.output.processFcns = {'removeconstantrows','mapminmax'};

% Setup Division of Data for Training, Validation, Testing
% For a list of all data division functions type: help nndivision
network.divideFcn = 'divideind';  % Devide Data set into Test and Train
network.divideMode = 'sample';  % Divide up every sample
% network.divideParam.trainRatio = 100/100;
% network.divideParam.valRatio = 0/100;
% network.divideParam.testRatio = 0/100;

Q = 130;
trainInd =1:100;
valInd = [];
testInd =101:130; 

[trainInd,valInd,testInd] = divideind(Q,trainInd,valInd,testInd); % Indizes für Test und TrainingsSet

% Validation: Patch15 + 18 --> Index 18 + Index 21 {"Patch15",-32.6529000000000,-86.0407000000000}
% Test: Vollbelegt --> Index 35 {"Vollbelegt01",-33.6531000000000,-89.8866000000000}
% net.divideFcn = 'divideind';  % Divide data according to index
% [trainInd,valInd,testInd] = divideind(36,[1:17, 19:20, 22:34, 36], [18, 21], 35);
% net.divideParam.trainInd = trainInd;
% net.divideParam.valInd   = valInd;
% net.divideParam.testInd  = testInd;

% Aktivierungsfunktion definieren
%net.layers{1}.transferfkn

% Choose a Performance Function
% For a list of all performance functions type: help nnperformance
network.performFcn = 'mse';  % Mean Squared Error

% Choose Plot Functions
% For a list of all plot functions type: help nnplot
network.plotFcns = {'plotperform','plottrainstate','ploterrhist', ...
    'plotregression', 'plotfit'};

% Train the Network
%[net,tr] = train(net,x,t);
%%Number of networks to train 
numNN = 100;
network_array = cell(1, numNN);
for i = 1:numNN
  fprintf('Training %d/%d\n', i, numNN)
  network_array{i} = train(network, x, t,'CheckpointFile','MyCheckpoint','CheckpointDelay',120);
  save 20212101_Workspace_NN_Torsion % Biegung und Torsion
end

save 20212101_Workspace_NN_Torsion

% Define Test Data, 101- 130
testdataset = t(101:130);

% Test the Network, but for more networks

errors = zeros (1, numNN);
perfs = zeros (1, numNN);
for i = 1:numNN
neti = network_array(i);
y2 = neti(testdataset);
perfs(i) = perform(network,t,y2);
errors(i) = gsubtract(t,y2);
save 20212101_Workspace_NN_Torsion
end

I trained 100 network and try to get their performance and errors. I extract the test data and it should be tested with the testdataset. It actually works.

But it always shows " Array indices must be positive integers or logical values" when it comes to code

y2 = neti(testdataset);

What could be the problem?

I appreciate any help!

Answer 1

[Salman Ahmed]

Hi Jingyuan Yao,

From my understanding, you have made the following typo in your code,

Change it from :

neti = network_array(i);

to :

neti = network_array{i};

For more information, refer to the page access data in cell array.