k-means python代码

k-means python代码

 

 

def fun_dis (x,y,n):
    return sum (map (lambda v1,v2:pow(abs(v1-v2),n), x,y))

def distance (x,y):
    #return fun_dis (x,y,2)
    return fun_dis (x,y,1)
    

def min_dis_center(center, node):
    min_index = 0
    min_dista = distance (center[0][1],node)
    for i in range (1,len(center)):
        tmp = distance (center[i][1],node)
        if (tmp < min_dista):
            min_dista = tmp
            min_index = i
    return min_index


# input [[x1,y1,z1..],[x2,y2,z2..]..]
# return 
def k_means (info,k=5):
    center = [[1,info[i]] for i in range(k)]
    result = [[i] for i in range(k)]
    width  = len (info[0])
    for i in range(k,len(info)):
        min_center = min_dis_center (center,info[i])
        for j in range(width):
            center[min_center][1][j] = (center[min_center][1][j] * center[min_center][0] + info[i][j])/ (1.0+center[min_center][0])
        center[min_center][0] += 1
        result[min_center].append (i)
    return result,center


test_data = [[1,100,300,9],
             [9,200,400,20],
             [2,102,305,10],
             [3,103,299,9],
             [3,105,299,11],
             [3,105,299,11],
             [9,200,400,20],
             [2,102,305,10],
             [11,220,409,20]]


data = k_means (test_data,2)
print data[0]
print data[1]