帮我写一下matalab代码，径向基函数神经网络，基于梯度下降法确定权值与中心，5折交叉验证

好的，以下是一个简单的径向基函数神经网络的MATLAB代码，包括权值和中心的确定以及5折交叉验证：

% 数据准备
load iris_dataset.mat
X = irisInputs';  % 输入数据
Y = irisTargets'; % 输出数据

% 神经网络参数
k = 10;  % 隐层神经元个数
lr = 0.1; % 学习率
epochs = 1000;  % 迭代次数

% 5折交叉验证
cv = cvpartition(size(X, 1), 'KFold', 5);
for i = 1:cv.NumTestSets
    trIdx = cv.training(i);
    teIdx = cv.test(i);
    X_train = X(trIdx, :);
    Y_train = Y(trIdx, :);
    X_test = X(teIdx, :);
    Y_test = Y(teIdx, :);

    % 确定中心
    centers = kmeans(X_train, k);

    % 计算径向基函数输出
    phi_train = zeros(size(X_train, 1), k);
    for j = 1:k
        phi_train(:, j) = exp(-sum((X_train - centers(j, :)).^2, 2));
    end
    phi_train = [ones(size(phi_train, 1), 1) phi_train];

    % 随机初始化权值
    W = randn(k + 1, size(Y_train, 2));

    % 训练神经网络
    for epoch = 1:epochs
        % 计算预测值
        Y_pred = phi_train * W;

        % 计算损失函数
        loss = mean(sum((Y_pred - Y_train).^2, 2));

        % 计算梯度
        grad = phi_train' * (Y_pred - Y_train) / size(X_train, 1);

        % 更新权值
        W = W - lr * grad;
    end

    % 计算测试集的径向基函数输出
    phi_test = zeros(size(X_test, 1), k);
    for j = 1:k
        phi_test(:, j) = exp(-sum((X_test - centers(j, :)).^2, 2));
    end
    phi_test = [ones(size(phi_test, 1), 1) phi_test];

    % 计算测试集的预测值
    Y_pred_test = phi_test * W;

    % 计算测试集的损失函数
    loss_test = mean(sum((Y_pred_test - Y_test).^2, 2));

    fprintf('Fold %d: training loss = %f, testing loss = %f\n', i, loss, loss_test);
end

希望对你有所帮助！