Possibilistic Fuzzy c-Means Clustering python代码

以下是Python实现的Possibilistic Fuzzy c-Means聚类算法代码：

import numpy as np
from sklearn.metrics.pairwise import euclidean_distances

def possibilistic_fuzzy_c_means(X, c, m, epsilon, max_iter):
    # 初始化隶属度矩阵U
    n, d = X.shape
    U = np.random.rand(n, c)
    U = np.divide(U, np.sum(U, axis=1)[:, np.newaxis])
    # 初始化聚类中心
    centers = np.random.rand(c, d)
    # 迭代更新
    for i in range(max_iter):
        # 更新隶属度矩阵U
        D = euclidean_distances(X, centers)
        U_new = np.power(np.divide(1, D), 2/(m-1))
        U_new = np.nan_to_num(U_new)
        U_new = np.divide(U_new, np.sum(U_new, axis=1)[:, np.newaxis])
        # 计算聚类中心
        centers_new = np.zeros((c, d))
        for j in range(c):
            centers_new[j, :] = np.sum(np.multiply(np.power(U_new[:, j], m), X), axis=0)
            centers_new[j, :] = np.divide(centers_new[j, :], np.sum(np.power(U_new[:, j], m)))
        # 判断是否收敛
        if np.linalg.norm(U - U_new) < epsilon:
            break
        U = U_new
        centers = centers_new
    return U, centers

其中，X为输入数据，c为聚类数，m为模糊度参数，epsilon为收敛阈值，max_iter为最大迭代次数。函数返回隶属度矩阵U和聚类中心centers。