K-Means implementation

kmeans.py

from random import randint

class KMeans:
    
    def __init__(self, minimum, maximum, dim, k):
        self._centroids = [[randint(minimum, maximum) for d in range(dim)] for x in range(k]
        self._points = {i: [] for i in range(k)}
        
    def update(self, vector):
        distance = lambda v: sqrt(sum([(x2-x1) ** 2 for x1, x2 in zip(v[1], vector)]))
        centroid_index = min(enumerate(self._centroids), key=dist)[0]
        self._points[centroid_index].append(vector)
        self._centroids[centroid_index] = [sum(xi) / len(xi) for xi in zip(*self._points[centroid_index])]