我嘗試比較sklearn包中和從頭開始的 kmean 聚類結果。暫存代碼如下所示：import matplotlib.pyplot as pltfrom matplotlib import stylestyle.use('ggplot')import numpy as npcolors = 10 * ["g", "r", "c", "b", "k"]class K_Means:    def __init__(self, k=3, tol=0.001, max_iter=300):        self.k = k        self.tol = tol        self.max_iter = max_iter    def fit(self, data):        self.centroids = {}        for i in range(self.k):            self.centroids[i] = data[i]        for i in range(self.max_iter):            self.classifications = {}            for i in range(self.k):                self.classifications[i] = []            for featureset in data:                distances = [np.linalg.norm(featureset - self.centroids[centroid]) for centroid in self.centroids]                classification = distances.index(min(distances))                self.classifications[classification].append(featureset)            prev_centroids = dict(self.centroids)            for classification in self.classifications:                self.centroids[classification] = np.average(self.classifications[classification], axis=0)            optimized = True            for c in self.centroids:                original_centroid = prev_centroids[c]                current_centroid = self.centroids[c]                if np.sum((current_centroid - original_centroid) / original_centroid * 100.0) > self.tol:                    print(np.sum((current_centroid - original_centroid) / original_centroid * 100.0))                    optimized = False            if optimized:                break    def predict(self, data):        distances = [np.linalg.norm(data - self.centroids[centroid]) for centroid in self.centroids]        classification = distances.index(min(distances))        return classification但由于收斂質心不同，結果也不同。sklearn 的散點圖：同時，上面代碼的散點圖：我想知道臨時代碼中存在哪些錯誤。