matlab svr代码

Sure! Here's an example of MATLAB code for Support Vector Regression (SVR):

% Load the dataset
load('data.mat'); % Replace 'data.mat' with the actual filename

% Split the dataset into training and testing sets
train_ratio = 0.8; % Adjust the ratio as needed
train_size = round(train_ratio * size(data, 1));
train_data = data(1:train_size, 1:end-1);
train_labels = data(1:train_size, end);
test_data = data(train_size+1:end, 1:end-1);
test_labels = data(train_size+1:end, end);

% Normalize the data if needed
train_data = zscore(train_data);
test_data = zscore(test_data);

% Train the SVR model
model = fitrsvm(train_data, train_labels, 'KernelFunction', 'gaussian');

% Predict on the test data
predictions = predict(model, test_data);

% Evaluate the model
mse = mean((predictions - test_labels).^2);
rmse = sqrt(mse);
r2 = 1 - sum((predictions - test_labels).^2) / sum((test_labels - mean(test_labels)).^2);

% Display the results
fprintf('Root Mean Squared Error (RMSE): %.4f\n', rmse);
fprintf('Coefficient of Determination (R^2): %.4f\n', r2);

Make sure to replace `'data.mat'` with the actual filename of your dataset. Also, you can adjust the parameters and kernel function in the `fitrsvm` function according to your requirements.

Remember to have the Statistics and Machine Learning Toolbox installed in MATLAB to use the SVR functions.