K_means一种确定k值的算法_也可与K_means++一起使用

 

#-*- coding:utf-8 -*-
#! /usr/bin/env python

from numpy import *

d1 = 3
d2 = 4
d3 = 5
d4 = 6
d5 = 100
d6 = 101
d7 = 102
d8 = 1001
d9 = 1002
d10 = 1003
s = [d1,d2,d3,d4,d5,d6,d7,d8,d9,d10]
def segment(s):
    print "s is %s"%s
    gap = 0
    #strip_gap_value = 5
    strip_gap_value1 = sum(s)/len(s)
    print "strip_gap_value1 is %s"%strip_gap_value1
    # print "sum(s) is %s"%sum(s)
    # print "len(s) is %s"%len(s)
    strip_gap_value2 = s[-1]-s[0]
    #strip_gap_value = min(strip_gap_value1,strip_gap_value2)
    strip_gap_value = strip_gap_value2/len(s)
    #strip_gap_value = 5
    print "strip_gap_value is %s"%strip_gap_value
    #1尝试用均值不可行2尝试用最大值与最小值差值，再与均值比较最小值不可行。3 最大值和最小值差值除以个数
    gap = strip_gap_value
    last = s[0]
    sub_b = [s[0]]
    bucket = [sub_b]
    for i in s[1:]:
        feet = i - last
        if feet > gap:
            sub_b = [i]
            bucket.append(sub_b)
            gap = strip_gap_value
        # else:
        #     sub_b.append(i)
        #     gap = gap + feet
        else:
            sub_b.append(i)
            gap = i - sub_b[0]
            gap = gap/2.0 if gap>strip_gap_value else strip_gap_value
        last = i

    print bucket
    return bucket

　　

数值分组最新版