用matlab写一个kmeans函数

好的，我可以为您提供一个简单的 MATLAB K-Means 函数。

function [cluster_idx, centroid] = kmeans(data, k)

% Initialize centroids randomly
centroid = datasample(data, k, 'Replace', false);

% Initialize variables
cluster_idx = zeros(size(data,1), 1);
sse = Inf;
iter = 0;

% Continue until SSE no longer decreases
while true

    % Assign each data point to the closest centroid
    for i = 1:size(data,1)
        distances = sum((centroid - data(i,:)).^2, 2);
        [~, idx] = min(distances);
        cluster_idx(i) = idx;
    end
    
    % Update centroids to the mean of their assigned data points
    for j = 1:k
        centroid(j,:) = mean(data(cluster_idx == j,:), 1);
    end
    
    % Compute the SSE
    sse_new = 0;
    for i = 1:k
        sse_new = sse_new + sum(sum((data(cluster_idx == i,:) - centroid(i,:)).^2));
    end
    
    % Check if SSE has converged
    if abs(sse_new - sse) < 1e-6 || iter >= 100
        break;
    end
    
    % Update SSE and iteration number
    sse = sse_new;
    iter = iter + 1;
end

这个函数采用数据矩阵和聚类数量作为输入，并返回每个数据点的聚类分配和聚类中心。算法使用随机初始化来选择初始聚类中心，并在 SSE 不再减小或达到最大迭代次数时停止。