python学习笔记 python实现k-means聚类

# -*- coding: utf-8 -*-
"""
Created on Thu Mar 16 14:52:58 2017

@author: Jarvis
"""
import tensorflow as tf
import numpy as np
import pandas as pd
import math
import random
from pandas import Series,DataFrame
def cal_dis(a,b):
    sum = 0
    for x,y in zip(a,b):
        sum = sum+(x-y)*(x-y)
    return math.sqrt(sum)

def is_same_series(a,b):
    
    for x,y in zip(a,b):
        if x != y:
            return False
    return True
def is_constant_vec(a,b):
    if len(a) == 0 or len(b) == 0:
        return False
    
    for x,y in zip(a,b):
        if not is_same_series(x,y):
            return False 
    
    return True 

def init_typeSet(init_set,type_num):
    for i in range(type_num):
        init_set.append([])
    return init_set

def Mindis_type_no(x,vecs,typ_num = 4):
    mindis = cal_dis(x,vecs[0])
    ans = 0
    for i in range(1,type_num):
        tmp  = cal_dis(x,vecs[i])
        if mindis > tmp:
            ans= i
            mindis = cal_dis(x,vecs[i])
    return ans

def reduce_mean(vecs):
    
    reduce_vec = vecs[0].copy()
    print(reduce_vec)
    for i in range(1,len(vecs)):
        print (vecs[i][5])
        print (reduce_vec[5])
        reduce_vec = reduce_vec+vecs[i]
    
        print (reduce_vec[5])
    
    reduce_vec = reduce_vec/len(vecs)
#    print(reduce_vec)
    return reduce_vec
def get_vecs(sets):
    vecs = []
    for i in sets:
        vecs.append(data.ix[i])
    return vecs

raw_data_file = pd.read_csv('NDVI_NDWI_all.csv',header = None,encoding = 'gbk')
data = (raw_data_file)
del data[0]
del data[1]
del data[2]

type_num = 4
init_type_vec = []#类质心
tmp_set = set([])
data_size = len(data)

while (len(tmp_set) < type_num):
     tmp_set.add(random.choice(range(data_size)))

for i in tmp_set:
    tmp = data.ix[i]
    init_type_vec.append(tmp)

pre_vec = []
#print (is_constant_vec(pre_vec,init_type_vec))

while( not is_constant_vec(pre_vec,init_type_vec)):
    type_set = []
    type_set = init_typeSet(type_set,type_num)
    for j in range(len(data)):
        tmp_type = Mindis_type_no(data.ix[j],init_type_vec)
        type_set[tmp_type].append(j)
        #type_set[tmp_type].append(data.ix[j])
    if(len(pre_vec) == 0):
        pre_vec = init_type_vec.copy()
    else:
        for i in range(type_num):
            pre_vec[i] = init_type_vec[i]
            need_cal_vecs = get_ves(type_set[i])
            init_type_vec[i] = reduce_mean(need_cal_vecs).copy()
for i in range(type_num):
    print('--------------------------------')
    print(type_set[i])
    
    
with open('output.txt',"w"):
    for i in range(type_num):
        print ("type %d"%i)
        for j in type_set[i]:
            print(j)
    
#print(reduce_mean(type_set))
#for i in range(type_num):
     #   pre_vec[i] = tf.reduce_mean()