第14章 聚类方法 - k均值聚类 - 实现MyKmeans类

# kmeans

class MyKmeans:
    def __init__(self, k, n=20):
        self.k = k
        self.n = n
        
    def fit(self, x, centers=None):
        # 第一步，随机选择 K 个点, 或者指定
        if centers is None:
            idx = np.random.randint(low=0, high=len(x), size=self.k)
            centers = x[idx]
        #print(centers)
        
        inters = 0
        while inters < self.n:
            #print(inters)
            #print(centers)
            points_set = {key: [] for key in range(self.k)}

            # 第二步，遍历所有点 P，将 P 放入最近的聚类中心的集合中
            for p in x:
                nearest_index = np.argmin(np.sum((centers - p) ** 2, axis=1) ** 0.5)
                points_set[nearest_index].append(p)

            # 第三步，遍历每一个点集，计算新的聚类中心
            for i_k in range(self.k):
                centers[i_k] = sum(points_set[i_k])/len(points_set[i_k])
                
            inters += 1

        
            
        return points_set, centers