完成K-means算法的代码实现(同时提交源代码)及数据(至少测试5个数据集,数据集来源建议采用UCI数据集)测试结果

完成K-means聚类算法的代码实现通常会涉及以下几个步骤：

1. 导入必要的库：例如NumPy用于数值计算，pandas处理数据，sklearn提供预处理工具。

import numpy as np
from sklearn.datasets import load_iris, make_blobs, load_digits, fetch_openml
import pandas as pd

2. 数据加载和预处理：
- UCI数据集：可以使用`load_iris()`、`make_blobs()`等函数。
- OpenML数据集：`fetch_openml(data_id=your_data_id)`，替换`your_data_id`为你需要的数据ID。

def load_and_preprocess_data(name):
    if name == 'iris':
        data = load_iris()
    elif name == 'blobs':
        data = make_blobs(n_samples=100, centers=3)
    # 更多数据集...
    else:
        data = fetch_openml(data_id=your_data_id, return_X_y=True)
    
    X, _ = data
    return X

# 测试数据集
data_sets = ['iris', 'blobs', 'digits', 'your_data_name', 'another_data_name']
test_datasets = [load_and_preprocess_data(ds) for ds in data_sets]

3. K-means算法的核心实现：

def kmeans(X, k, max_iter=300):
    n_samples, n_features = X.shape
    centroids = init_centroids(X, k)  # 初始化质心
    labels = None
    
    for _ in range(max_iter):
        # 分配样本到最近的质心
        labels = assign_clusters(X, centroids)
        
        # 更新质心
        new_centroids = update_centroids(X, labels, k)
        
        # 检查是否收敛
        if np.allclose(centroids, new_centroids):
            break
        centroids = new_centroids
    
    return labels, centroids

# 初始化质心方法（这里简化了）
def init_centroids(X, k):
    rand_indices = np.random.choice(len(X), k, replace=False)
    centroids = X[rand_indices]
    return centroids

# 调整质心的方法
def update_centroids(X, labels, k):
    cluster_counts = np.bincount(labels, minlength=k)
    new_centroids = []
    for i in range(k):
        cluster_points = X[labels == i]
        new_centroids.append(np.mean(cluster_points, axis=0))
    return np.array(new_centroids)

# 转换簇标签到类别
def assign_clusters(X, centroids):
    distances = np.linalg.norm(X[:, np.newaxis] - centroids, axis=-1)
    return np.argmin(distances, axis=1)

4. 测试并输出结果：

for dataset in test_datasets:
    labels, centroids = kmeans(dataset, k=3)
    print(f"Dataset: {dataset.name}, Clusters: {list(range(1, k+1))}")
    print("Labels:", labels[:10])  # 输出前几个样本的簇标签
    print("Centroids:", centroids)