knn算法

自己造数据的能力是有限的，咱们当然可以借助外来的数据，进行练习，本次的数据源https://github.com/Jack-Cherish/Machine-Learning/blob/master/kNN/2.%E6%B5%B7%E4%BC%A6%E7%BA%A6%E4%BC%9A/datingTestSet.txt可以下载下来试用

import matplotlib.pyplot as plt
import matplotlib.colors
import  operator
import numpy as  np
#读取文件数据
def fileread(filem):
    f_open=open(filem)
    f_read=f_open.readlines()
    f_hang=len(f_read)
    dataH=np.zeros((f_hang,3))
    lable=[]
    index=0
    ########对数据进行处理，使得数据符合我们的分类器算法的要求
    for ha in f_read:
        f_qukong=ha.strip()#
        f_zhibaio=f_qukong.split('\t')#
        dataH[index,:]=f_zhibaio[:3]
        index+=1
        if f_zhibaio[-1]=="didntLike":
            lable.append(1)
        if f_zhibaio[-1] == "largeDoses":
            lable.append(2)
        if f_zhibaio[-1] == "smallDoses":
            lable.append(3)
    return dataH,lable
#构建分类器
def classfen(test, train, labe, k):
    global so
    hangshu = train.shape[0]
    xiangjian = np.tile(test, (hangshu, 1)) - train
    pingfang = xiangjian ** 2
    qiuhe = pingfang.sum(axis=1)
    kaifang = qiuhe ** 0.5
    paixu = kaifang.argsort()
    count = {}
    for i in range(k):
        lable = labe[paixu[i]]
        count[lable] = count.get(lable, 0) + 1
        so = sorted(count.items(), key=operator.itemgetter(1), reverse=True)
    return so[0][0]

if __name__ == '__main__':
    file='data.txt'
    gro, ll = fileread(file)
    te = [60000, 5,800]
    plt.scatter(gro[:,0],gro[:,1],c=ll)
    plt.scatter(te[0:1],te[1:2],marker='*',c='red',s=400)
    plt.show()
    tw = classfen(te, gro, ll, 3)
    print(te)
    print(te[0:1])

最近较忙，如有一问，欢迎博客留言