Python手动实现kmeans聚类和调用sklearn实现

1. 算法步骤

1. 随机选取k个样本点充当k个簇的中心点；
2. 计算所有样本点与各个簇中心之间的距离，然后把样本点划入最近的簇中；
3. 根据簇中已有的样本点，重新计算簇中心；
4. 重复步骤2和3，直到簇中心不再改变或改变很小。

2. 手动Python实现

?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

import numpy as np import matplotlib.pyplot as plt from sklearn.datasets.samples_generator import make_blobs   n_data = 400 n_cluster = 4 # generate training data X, y = make_blobs(n_samples=n_data, centers=n_cluster, cluster_std=0.60, random_state=0)   # generate centers of clusters centers = np.random.rand(4, 2)*5   EPOCH = 10 tol = 1e-5 for epoch in range(EPOCH):     labels = np.zeros(n_data, dtype=np.int)       # 计算每个点到簇中心的距离并分配label     for i in range(n_data):         distance = np.sum(np.square(X[i]-centers), axis=1)         label = np.argmin(distance)         labels[i] = label       # 重新计算簇中心     for i in range(n_cluster):         indices = np.where(labels == i)[0]       # 找出第i簇的样本点的下标         points = X[indices]         centers[i, :] = np.mean(points, axis=0)  # 更新第i簇的簇中心   plt.scatter(X[:, 0], X[:, 1], c=labels, s=40, cmap='viridis') plt.show()

运行结果：(注：当簇中心初始化不好时，可能计算会有点错误)

 

3. 调用sklearn实现kmeans

?

1 2 3 4 5 6 7 8 9 10 11 12 13

import matplotlib.pyplot as plt from sklearn.cluster import KMeans from sklearn.datasets.samples_generator import make_blobs   # Generate some data X, y = make_blobs(n_samples=400, centers=4, cluster_std=0.60, random_state=0)   # kmeans clustering kmeans = KMeans(4, random_state=0) kmeans.fit(X)   # 训练模型 labels = kmeans.predict(X)   # 预测分类 plt.scatter(X[:, 0], X[:, 1], c=labels, s=40, cmap='viridis') plt.show()

运行结果：