写一段k-means聚类的python代码

好的，以下是一个简单的 k-means 聚类的 Python 代码，用于将数据集分为 K 个簇：

import numpy as np

# 定义距离函数
def euclidean_distance(x1, x2):
    return np.sqrt(np.sum((x1 - x2) ** 2))

class KMeans:
    def __init__(self, k=5, max_iters=100, plot_steps=False):
        self.k = k
        self.max_iters = max_iters
        self.plot_steps = plot_steps

        # 初始化聚类中心
        self.centroids = None

    def _initialize_centroids(self, X):
        # 随机选择 K 个数据点作为初始聚类中心
        n_samples, n_features = X.shape
        self.centroids = X[np.random.choice(range(n_samples), self.k, replace=False)]

    def _create_clusters(self, X):
        # 将样本点分配到最近的聚类中心
        clusters = [[] for _ in range(self.k)]
        for idx, sample in enumerate(X):
            centroid_idx = self._closest_centroid(sample)
            clusters[centroid_idx].append(idx)
        return clusters

    def _closest_centroid(self, sample):
        # 计算样本点到各聚类中心的距离，返回距离最近的聚类中心的索引
        distances = [euclidean_distance(sample, point) for point in self.centroids]
        closest_idx = np.argmin(distances)
        return closest_idx

    def _calculate_centroids(self, X, clusters):
        # 计算每个簇的平均值，更新聚类中心
        n_features = X.shape[1]
        centroids = np.zeros((self.k, n_features))
        for idx, cluster in enumerate(clusters):
            cluster_mean = np.mean(X[cluster], axis=0)
            centroids[idx] = cluster_mean
        self.centroids = centroids

    def _get_cluster_labels(self, clusters, X):
        # 返回每个样本所属的簇的索引
        y_pred = np.zeros(X.shape[0])
        for cluster_idx, cluster in enumerate(clusters):
            for sample_idx in cluster:
                y_pred[sample_idx] = cluster_idx
        return y_pred

    def predict(self, X):
        # 进行 KMeans 聚类
        self._initialize_centroids(X)

        for i in range(self.max_iters):
            clusters = self._create_clusters(X)

            if self.plot_steps:
                self.plot(X, self.centroids, clusters)

            prev_centroids = self.centroids

            self._calculate_centroids(X, clusters)

            if self._is_converged(prev_centroids, self.centroids):
                break

        clusters = self._create_clusters(X)

        if self.plot_steps:
            self.plot(X, self.centroids, clusters)

        return self._get_cluster_labels(clusters, X)

    def _is_converged(self, prev_centroids, centroids):
        # 判断聚类中心是否发生变化
        return np.array_equal(prev_centroids, centroids)

使用示例：

from sklearn.datasets import make_blobs
import matplotlib.pyplot as plt

# 生成随机数据
X, y = make_blobs(centers=3, n_samples=500, random_state=1)

# 创建 KMeans 对象
kmeans = KMeans(k=3, max_iters=100, plot_steps=True)

# 进行聚类
y_pred = kmeans.predict(X)

# 显示聚类结果
plt.scatter(X[:, 0], X[:, 1], c=y_pred)
plt.show()