Hello, i am trying to code an automatic detection of alzheimer from EEG signals but my code has an error when using trainNetwork. It worked perfectely with a SVM but doesn’t with a CNN. I tried looking online but nothing seems too work. I got this error :
Error using trainNetwork (line 191)
Too many input arguments.
Error in CNN (line 178)
net = trainNetwork(X_train, y_train, layers, options);
Caused by:
Error using gather
Too many input arguments.
Does anyone have an idea. Here is the part of my code that produce the CNN :
X = all_features{:, 1:end-1}; % Use parentheses () for table indexing
y = all_features.Label;

y = categorical(y);

disp([‘Feature matrix dimensions: ‘, num2str(size(X))]);
disp([‘Labels vector dimensions: ‘, num2str(size(y))]);

X = zscore(X);

numFeatures = size(X, 2);
numObservations = size(X, 1);
X = reshape(X, [numObservations, numFeatures, 1, 1]); % Reshape for CNN

layers = [
imageInputLayer([numFeatures 1 1])
convolution2dLayer([3 1], 8, ‘Padding’, ‘same’)
batchNormalizationLayer
reluLayer
maxPooling2dLayer([2 1], ‘Stride’, 2)
convolution2dLayer([3 1], 16, ‘Padding’, ‘same’)
batchNormalizationLayer
reluLayer
fullyConnectedLayer(2)
softmaxLayer
classificationLayer];

options = trainingOptions(‘adam’, …
‘MaxEpochs’, 30, …
‘MiniBatchSize’, 16, …
‘InitialLearnRate’, 0.001, …
‘ValidationFrequency’, 10, …
‘Verbose’, false, …
‘Plots’, ‘training-progress’);
% 5-Fold Cross-Validation
cv = cvpartition(y, ‘KFold’, 5, ‘Stratify’, true);

accuracies = zeros(cv.NumTestSets, 1);
confusion_matrices = cell(cv.NumTestSets, 1);

for k = 1:cv.NumTestSets

train_idx = training(cv, k);
test_idx = test(cv, k);

X_train = X(train_idx, :, :, :);
y_train = y(train_idx);
X_test = X(test_idx, :, :, :);
y_test = y(test_idx);

net = trainNetwork(X_train, y_train, layers, options);

y_pred = classify(net, X_test);

confusion_matrices{k} = confusionmat(y_test, y_pred);
cm = confusion_matrices{k};

accuracies(k) = sum(diag(cm)) / sum(cm(:));
end

mean_accuracy = mean(accuracies);
fprintf(‘Mean Accuracy across 5 folds: %.2f%%n’, mean_accuracy * 100);

save(‘eeg_cnn_classifier_cv.mat’, ‘net’);

disp(‘Confusion Matrices for each fold:’);
for k = 1:cv.NumTestSets
disp([‘Fold ‘, num2str(k), ‘:’]);
disp(confusion_matrices{k});
endHello, i am trying to code an automatic detection of alzheimer from EEG signals but my code has an error when using trainNetwork. It worked perfectely with a SVM but doesn’t with a CNN. I tried looking online but nothing seems too work. I got this error :
Error using trainNetwork (line 191)
Too many input arguments.
Error in CNN (line 178)
net = trainNetwork(X_train, y_train, layers, options);
Caused by:
Error using gather
Too many input arguments.
Does anyone have an idea. Here is the part of my code that produce the CNN :
X = all_features{:, 1:end-1}; % Use parentheses () for table indexing
y = all_features.Label;

y = categorical(y);

disp([‘Feature matrix dimensions: ‘, num2str(size(X))]);
disp([‘Labels vector dimensions: ‘, num2str(size(y))]);

X = zscore(X);

numFeatures = size(X, 2);
numObservations = size(X, 1);
X = reshape(X, [numObservations, numFeatures, 1, 1]); % Reshape for CNN

layers = [
imageInputLayer([numFeatures 1 1])
convolution2dLayer([3 1], 8, ‘Padding’, ‘same’)
batchNormalizationLayer
reluLayer
maxPooling2dLayer([2 1], ‘Stride’, 2)
convolution2dLayer([3 1], 16, ‘Padding’, ‘same’)
batchNormalizationLayer
reluLayer
fullyConnectedLayer(2)
softmaxLayer
classificationLayer];

options = trainingOptions(‘adam’, …
‘MaxEpochs’, 30, …
‘MiniBatchSize’, 16, …
‘InitialLearnRate’, 0.001, …
‘ValidationFrequency’, 10, …
‘Verbose’, false, …
‘Plots’, ‘training-progress’);
% 5-Fold Cross-Validation
cv = cvpartition(y, ‘KFold’, 5, ‘Stratify’, true);

accuracies = zeros(cv.NumTestSets, 1);
confusion_matrices = cell(cv.NumTestSets, 1);

for k = 1:cv.NumTestSets

train_idx = training(cv, k);
test_idx = test(cv, k);

X_train = X(train_idx, :, :, :);
y_train = y(train_idx);
X_test = X(test_idx, :, :, :);
y_test = y(test_idx);

net = trainNetwork(X_train, y_train, layers, options);

y_pred = classify(net, X_test);

confusion_matrices{k} = confusionmat(y_test, y_pred);
cm = confusion_matrices{k};

accuracies(k) = sum(diag(cm)) / sum(cm(:));
end

mean_accuracy = mean(accuracies);
fprintf(‘Mean Accuracy across 5 folds: %.2f%%n’, mean_accuracy * 100);

save(‘eeg_cnn_classifier_cv.mat’, ‘net’);

disp(‘Confusion Matrices for each fold:’);
for k = 1:cv.NumTestSets
disp([‘Fold ‘, num2str(k), ‘:’]);
disp(confusion_matrices{k});
end Hello, i am trying to code an automatic detection of alzheimer from EEG signals but my code has an error when using trainNetwork. It worked perfectely with a SVM but doesn’t with a CNN. I tried looking online but nothing seems too work. I got this error :
Error using trainNetwork (line 191)
Too many input arguments.
Error in CNN (line 178)
net = trainNetwork(X_train, y_train, layers, options);
Caused by:
Error using gather
Too many input arguments.
Does anyone have an idea. Here is the part of my code that produce the CNN :
X = all_features{:, 1:end-1}; % Use parentheses () for table indexing
y = all_features.Label;

y = categorical(y);

disp([‘Feature matrix dimensions: ‘, num2str(size(X))]);
disp([‘Labels vector dimensions: ‘, num2str(size(y))]);

X = zscore(X);

numFeatures = size(X, 2);
numObservations = size(X, 1);
X = reshape(X, [numObservations, numFeatures, 1, 1]); % Reshape for CNN

layers = [
imageInputLayer([numFeatures 1 1])
convolution2dLayer([3 1], 8, ‘Padding’, ‘same’)
batchNormalizationLayer
reluLayer
maxPooling2dLayer([2 1], ‘Stride’, 2)
convolution2dLayer([3 1], 16, ‘Padding’, ‘same’)
batchNormalizationLayer
reluLayer
fullyConnectedLayer(2)
softmaxLayer
classificationLayer];

options = trainingOptions(‘adam’, …
‘MaxEpochs’, 30, …
‘MiniBatchSize’, 16, …
‘InitialLearnRate’, 0.001, …
‘ValidationFrequency’, 10, …
‘Verbose’, false, …
‘Plots’, ‘training-progress’);
% 5-Fold Cross-Validation
cv = cvpartition(y, ‘KFold’, 5, ‘Stratify’, true);

accuracies = zeros(cv.NumTestSets, 1);
confusion_matrices = cell(cv.NumTestSets, 1);

for k = 1:cv.NumTestSets

train_idx = training(cv, k);
test_idx = test(cv, k);

X_train = X(train_idx, :, :, :);
y_train = y(train_idx);
X_test = X(test_idx, :, :, :);
y_test = y(test_idx);

net = trainNetwork(X_train, y_train, layers, options);

y_pred = classify(net, X_test);

confusion_matrices{k} = confusionmat(y_test, y_pred);
cm = confusion_matrices{k};

accuracies(k) = sum(diag(cm)) / sum(cm(:));
end

mean_accuracy = mean(accuracies);
fprintf(‘Mean Accuracy across 5 folds: %.2f%%n’, mean_accuracy * 100);

save(‘eeg_cnn_classifier_cv.mat’, ‘net’);

disp(‘Confusion Matrices for each fold:’);
for k = 1:cv.NumTestSets
disp([‘Fold ‘, num2str(k), ‘:’]);
disp(confusion_matrices{k});
end cnn, machine learning, deep learning MATLAB Answers — New Questions

​

I have this code and i want to save the variable x from workspace into a file clc;
clear;
close all;
% Parameters
sampling_frequency = 1000; % Sampling frequency in Hz
duration = 5; % Duration of signal acquisition in seconds
num_samples = sampling_frequency * duration;
% Initialize variables
time = (0:num_samples-1) / sampling_frequency;
x = zeros(1, num_samples);
% Connect to Arduino
a = arduino();
% Bandpass filter parameters
low_cutoff_frequency = 6; % Low cutoff frequency in Hz
high_cutoff_frequency = 400; % High cutoff frequency in Hz
% Design the bandpass filter
[b, a_filter] = butter(2, [low_cutoff_frequency, high_cutoff_frequency] / (sampling_frequency / 2), ‘bandpass’); %Normalized between [0 1]
% Plot for real-time visualization of raw signal
figure(‘Name’, ‘Real-Time Raw EMG Signal’);
h_raw = plot(nan, nan); % Initialize plot with NaN
grid on;
title(‘Real-Time Raw EMG Signal’);
xlabel(‘Time (s)’);
ylabel(‘Voltage (V)’);
xlim([0, duration]); % Adjust xlim as needed
ylim([0, 5]); % Adjust ylim according to expected voltage range
% Plot for real-time visualization of filtered signal
figure(‘Name’, ‘Real-Time Filtered EMG Signal’);
h_filtered = plot(nan, nan); % Initialize plot with NaN
grid on;
title(‘Real-Time Filtered EMG Signal’);
xlabel(‘Time (s)’);
ylabel(‘Voltage (V)’);
xlim([0, duration]); % Adjust xlim as needed
ylim([0, 5]); % Adjust ylim according to expected voltage range
% Start loop to acquire signal
for i = 1:num_samples
% Read voltage
raw_voltage = readVoltage(a, ‘A0’);

% Store raw voltage
x(i) = raw_voltage;

% Apply bandpass filter to the signal up to current point if enough data points are collected
if i >= 25
filtered_signal = filtfilt(b, a_filter, x(1:i));

% Remove DC offset by subtracting the mean of the filtered signal
filtered_signal = filtered_signal – mean(filtered_signal);

% Update plot for filtered signal
set(h_filtered, ‘YData’, filtered_signal, ‘XData’, time(1:i));
end

% Update plot for raw signal
set(h_raw, ‘YData’, x(1:i), ‘XData’, time(1:i));

drawnow;

end
save(‘EMG_data.mat’, ‘x’,);
It doesn’t create the .mat file and I suppose the problem is that i stop the iteration very early. But I want to save any values that are obtained until any iteration. How can I fix that?I have this code and i want to save the variable x from workspace into a file clc;
clear;
close all;
% Parameters
sampling_frequency = 1000; % Sampling frequency in Hz
duration = 5; % Duration of signal acquisition in seconds
num_samples = sampling_frequency * duration;
% Initialize variables
time = (0:num_samples-1) / sampling_frequency;
x = zeros(1, num_samples);
% Connect to Arduino
a = arduino();
% Bandpass filter parameters
low_cutoff_frequency = 6; % Low cutoff frequency in Hz
high_cutoff_frequency = 400; % High cutoff frequency in Hz
% Design the bandpass filter
[b, a_filter] = butter(2, [low_cutoff_frequency, high_cutoff_frequency] / (sampling_frequency / 2), ‘bandpass’); %Normalized between [0 1]
% Plot for real-time visualization of raw signal
figure(‘Name’, ‘Real-Time Raw EMG Signal’);
h_raw = plot(nan, nan); % Initialize plot with NaN
grid on;
title(‘Real-Time Raw EMG Signal’);
xlabel(‘Time (s)’);
ylabel(‘Voltage (V)’);
xlim([0, duration]); % Adjust xlim as needed
ylim([0, 5]); % Adjust ylim according to expected voltage range
% Plot for real-time visualization of filtered signal
figure(‘Name’, ‘Real-Time Filtered EMG Signal’);
h_filtered = plot(nan, nan); % Initialize plot with NaN
grid on;
title(‘Real-Time Filtered EMG Signal’);
xlabel(‘Time (s)’);
ylabel(‘Voltage (V)’);
xlim([0, duration]); % Adjust xlim as needed
ylim([0, 5]); % Adjust ylim according to expected voltage range
% Start loop to acquire signal
for i = 1:num_samples
% Read voltage
raw_voltage = readVoltage(a, ‘A0’);

% Store raw voltage
x(i) = raw_voltage;

% Apply bandpass filter to the signal up to current point if enough data points are collected
if i >= 25
filtered_signal = filtfilt(b, a_filter, x(1:i));

% Remove DC offset by subtracting the mean of the filtered signal
filtered_signal = filtered_signal – mean(filtered_signal);

% Update plot for filtered signal
set(h_filtered, ‘YData’, filtered_signal, ‘XData’, time(1:i));
end

% Update plot for raw signal
set(h_raw, ‘YData’, x(1:i), ‘XData’, time(1:i));

drawnow;

end
save(‘EMG_data.mat’, ‘x’,);
It doesn’t create the .mat file and I suppose the problem is that i stop the iteration very early. But I want to save any values that are obtained until any iteration. How can I fix that? I have this code and i want to save the variable x from workspace into a file clc;
clear;
close all;
% Parameters
sampling_frequency = 1000; % Sampling frequency in Hz
duration = 5; % Duration of signal acquisition in seconds
num_samples = sampling_frequency * duration;
% Initialize variables
time = (0:num_samples-1) / sampling_frequency;
x = zeros(1, num_samples);
% Connect to Arduino
a = arduino();
% Bandpass filter parameters
low_cutoff_frequency = 6; % Low cutoff frequency in Hz
high_cutoff_frequency = 400; % High cutoff frequency in Hz
% Design the bandpass filter
[b, a_filter] = butter(2, [low_cutoff_frequency, high_cutoff_frequency] / (sampling_frequency / 2), ‘bandpass’); %Normalized between [0 1]
% Plot for real-time visualization of raw signal
figure(‘Name’, ‘Real-Time Raw EMG Signal’);
h_raw = plot(nan, nan); % Initialize plot with NaN
grid on;
title(‘Real-Time Raw EMG Signal’);
xlabel(‘Time (s)’);
ylabel(‘Voltage (V)’);
xlim([0, duration]); % Adjust xlim as needed
ylim([0, 5]); % Adjust ylim according to expected voltage range
% Plot for real-time visualization of filtered signal
figure(‘Name’, ‘Real-Time Filtered EMG Signal’);
h_filtered = plot(nan, nan); % Initialize plot with NaN
grid on;
title(‘Real-Time Filtered EMG Signal’);
xlabel(‘Time (s)’);
ylabel(‘Voltage (V)’);
xlim([0, duration]); % Adjust xlim as needed
ylim([0, 5]); % Adjust ylim according to expected voltage range
% Start loop to acquire signal
for i = 1:num_samples
% Read voltage
raw_voltage = readVoltage(a, ‘A0’);

% Store raw voltage
x(i) = raw_voltage;

% Apply bandpass filter to the signal up to current point if enough data points are collected
if i >= 25
filtered_signal = filtfilt(b, a_filter, x(1:i));

% Remove DC offset by subtracting the mean of the filtered signal
filtered_signal = filtered_signal – mean(filtered_signal);

% Update plot for filtered signal
set(h_filtered, ‘YData’, filtered_signal, ‘XData’, time(1:i));
end

% Update plot for raw signal
set(h_raw, ‘YData’, x(1:i), ‘XData’, time(1:i));

drawnow;

end
save(‘EMG_data.mat’, ‘x’,);
It doesn’t create the .mat file and I suppose the problem is that i stop the iteration very early. But I want to save any values that are obtained until any iteration. How can I fix that? save variable from workspace, iterations MATLAB Answers — New Questions

​