Python实现KNN算法

#coding:utf-8
import requests, json, time, re, os, sys, time
import urllib2
import random
import numpy as np

#设置为utf-8模式
reload(sys)
sys.setdefaultencoding( "utf-8" )


#读取文本文件,构建二维数组
def readDataFile(filename,format):
    if format:
        pass
    else:
        format = ‘,‘
    list = []
    #去除首位空格
    filename = filename.strip()
    #判断数据文件是否存在
    if os.path.isfile(filename):
        pass
        file_object = open(filename,‘rb‘)
        lines = file_object.readlines()
        for line in lines:
            tmp = []
            line = line.strip()
            for value in line.split(format)[:-1]:
                tmp.append(float(value))
            tmp.append(line.split(format)[-1])
            list.append(tmp)
    else:
        print "%s is not exists " % (filename)
    return list

#读取文本数据,拆分原始数据为特征和标签,返回特征值和标签值
def createData(filename,format=‘,‘):
    data_label = readDataFile(filename,format)
    if len(data_label) > 0:
        label = []
        data = []
        #data_label = [[1,100,123,‘A‘],[2,99,123,‘A‘],[100,1,12,‘B‘],[99,2,23,‘B‘]]
        for each in data_label:
            label.append(each[-1])
            data.append(each[:-1])
        return data,label

#根据输入数据和测试数据,进行分类
def calculateDistance(input,data,label,k):
    classes = ‘Error‘
    
    if len(data[0])==0 or len(label) == 0:
        print ‘data or label is null‘
        pass 
    elif k > len(data) :
        print "k : %s is out of bounds" % (k)
        pass 
    elif len(input) <> len(data[0]):
        print "特征变量值不够,输入变量特征个数为：%s,训练特征变量个数为：%s" % (len(input),len(data[0]))
        pass 
    else:
        result = []
        length = len(input)
        for i in range(len(data)):
            sum = 0
            for j in range(length):
                #pow(5,2) 标识5的平方为25,取两点之间的距离的平方并累加
                sum = sum + pow(input[j] - data[i][j],2)
            #取平方根
            sum = pow(sum,0.5)
            result.append(sum)
        #print result
        result = np.array(result)
        
        #argsort()根据元素的值从小到大对元素进行排序，返回下标
        sortedDistIndex = np.argsort(result)
        
        #统计前k个数中各个标签的个数
        classCount={}
        for i in range(k):
            voteLabel = label[sortedDistIndex[i]]
            ###对选取的K个样本所属的类别个数进行统计
            #dict.get(key, default=None) 返回指定键的值，如果值不在字典中返回默认值None。
            classCount[voteLabel] = classCount.get(voteLabel,0) + 1
        ###选取出现的类别次数最多的类别
        maxCount = 0
        for key,value in classCount.items():
            if value > maxCount:
                maxCount = value
                classes = key
    return classes 

filename = ‘/home/shutong/jim/crawl/data.csv‘
data,label = createData(filename)
input = [1,20]
k = 4
result = calculateDistance(input,data,label,k)
print input,result