Probleme de tracer l'évolution de l'erreur quadratique moyenne des algorithmes d'apprentissage automatique.

bonjour, svp comment faire pour tracer l'evolution de l'erreur quadratique moyenne pour les diffrents algorithmes d'apprentissage automatique tels que :LASSO,ELASTICNET, BAGTREE,BOOST,SVR, ELM, KRR et GPR.voila ce que j'ai fait mais c'est nul je ne comprend pas.

% Import the CSV data

X_train = readtable('traindata.csv');

% Display the names of the variables in the dataset

variableNames = X_train.Properties.VariableNames;

fprintf('The names of the variables in the dataset are:\n');

for i = 1:length(variableNames)

fprintf('%s\n', variableNames{i});

end

who;

% Get the values of the `I1` variable

I1=X_train.I1;

I2=X_train.I2;

I3=X_train.I3;

Vdcmean1=X_train.Vdcmean1;

Pdcmean1=X_train.Pdcmean1;

% Select the `I1`, `I2`, and `i3` variables

I1I2I3= X_train(:, {'I1', 'I2', 'I3' 'Pdcmean1'});

% Créez un tableau contenant les noms des variables à utiliser pour les algorithmes de processus gaussiens

variable_names = {'I1', 'I2', 'I3', 'Pdcmean1'};

% Créez un tableau contenant les données à utiliser pour les algorithmes de processus gaussiens

X_train = X_train(:, variable_names);

% Créez un vecteur contenant les labels de sortie

labels = X_train.Pdcmean1;

% Itérez sur les différents algorithmes de processus gaussiens

function_mean=@(x)0;

function_cov=@(x,x_prime)exp-(x-x_prime)*(x-x_prime)/0.1;

x=X_train;

gp=GaussienProcessRegressor(function_mean, function_cov);

algorithm_name = {'GaussianProcessRegression', 'GaussianProcessClassification'};

gp = GaussianProcessRegressor(function_mean, function_cov);

% Créez un objet algorithme

algorithm = feval(algorithm_name, data, labels);

% Entraînez l'algorithme

[algorithm, train_rmse] = train(algorithm, data);

% Tracez l'évolution du RMSE lors de la phase de formation

figure;

plot(train_rmse);

xlabel('Étape d''entraînement');

ylabel('RMSE');

title(algorithm_name);

%Separation des variable et des classes

%variables=data(:, 1:10);

%classe =data(:, 1:3);

% Separation of the features and label

%X_train=data(:,{'I1','I2','I1MAX','I1MIN','I1VAR','I2MAX','I2MIN','I2VAR','I3','I4','I3max','I3min','I3var','I4MAX','I4MIN','I5','I6','Itotal1','Pdcmean1','class'});

% Split the data into training and test sets

X_train = data(1:600, 1:30);

%y_train = data(1:600, 31);

X_test = data(101:end, 1:2);

%y_test = data(101:end, 3);

% Create a list of regression models

%Convert categorical variables to dummy variables

% Train the GPR model

model_gpr = fitrgp(X_train, y_train);

% Train the LASSO model

model_lasso = lasso(X_train, y_train);

% Train the Elastic net model

model_elastic_net = elasticnet(X_train, y_train);

% Train the Bagging model

model_bagging = bagging(X_train, y_train);

% Train the boosting model

model_boosting = boosting(X_train, y_train);

% Train the KRR model

model_krr = kernel_ridge(X_train, y_train);

% Train the SVR model

model_svr = support_vector_regression(X_train, y_train);

% Train the ELM model

model_elm = extreme_learning_machine(X_train, y_train);

% Train the RVM model

model_rvm = relevance_vector_machine(X_train, y_train);

% Initialize the RMSEs

rmses = zeros(1, length(models));

% Train the models

for i = 1:length(models)

% Train the model

model = eval(models{i});

model.fit(X_train, y_train);

% Calculate the RMSE on the training set

y_pred = model.predict(X_train);

(i, 1) = sqrt(mean((y_pred - y_train).^2));

Calculate the RMSE on the test set

= model.predict(X_test);

vrmses(i, 2) = sqrt(mean((y_pred - y_test).^2));

end

Plot the RMSEs

figure;

plot(rmses(:, 1), '-o', 'MarkerSize', 10, 'LineWidth', 2);

hold on;

plot(rmses(:, 2), '-x', 'MarkerSize', 10, 'LineWidth', 2);

legend('RMSE sur données d''entrainement', 'RMSE sur données de test');

xlabel('etape d''entrainement');

ylabel('RMSE');