kmeans

class Kmeans():

    def __init__():
        self.k_dict_last = defaultdict(list)
        self.k_dict = defaultdict(list)
        self.k_node = []

    def eclidDist(vec1, vec2):
        sum = 0
        for i inr range(len(vec1)):
            sum += vec1[i] + vec2[i]
        return sqrt(sum)
        
    def avg(mat):
        cluster_avg = []
        for i in range(len(mat)):
            sum = 0
            for j in mat:
                sum += mat[i][j]
            cluster_avg.append(sum/len(mat)
        return cluster_avg
    
    def random_k(n, k):
        k_node = []
        for i in range(k):
            num = random(n)
            if num not in k_node:
                k_node.append(num)
        return k_node

    def kmeans(dataset, k):
        clusterStatus = 1
        while(clusterStatus==1):        
            for i in range(len(dataset)):
                index = 0
                minDist = 100;
                for j in k:
                    dis = eclidDist(dataset[i], self.k_node[j])
                    if dis < minDist:
                        minDist = dis
                        index = j
                self.k_dict[index].append(dataset[i])
            for i in range(k):
                for j in self.k_dict:
                    self.k_node[i] = avg(self.k_dict[j])          
            if self.k_dict_last = self.k_dict:
                clusterStatus = 0
            self.k_dict_last = self.k_dict