机器学习实战-k近邻分类

k-近邻算法（KNN）

一。工作原理

存在一个样本数据集合，即训练样本集，并且样本集中每个数据都存在标签（样本集中每一数据与所属分类的对应关系），

输入没有标签的新数据后，将新数据的每个特征与样本集中数据对应的特征进行比较，然后算法提取样本集中特征最相似数据的分类标签，

一般只选择样本数据集中前k个最相似结果，通常k是不大于20的整数，将k个最相似数据中出现次数最多的分类，作为新数据的分类。

二。一般流程

1.收集数据

2.准备数据，计算所需要的数值

3.分析数据

4.训练数据

5.测试数据

6.使用算法：首先输入样本数据和结构化的输出结果，然后运行k-近邻算法判定输入数据分别属于哪个分类，最后应用

对计算出的分类执行后续的处理。

 三。代码

<pre name="code" class="python">from numpy import *
import operator
import sys
import matplotlib
import matplotlib.pyplot as plt

def createDataSet() :
        group = array([[1.0,1.1],[1.0,1.0],[0,0],[0,0.1]])
        labels = ['A','A','B','B']
        return group, labels

def file2matrix(filename) :
        fr = open(filename)
        arrayOLines = fr.readlines()
        #get the line numbers
        numberOfLines = len(arrayOLines)
        returnMat = zeros((numberOfLines, 3))
        classLabelVector = []

        idx=0
        for line in arrayOLines :
                line = line.strip()
                listFromLine = line.split('\t')
                returnMat[idx:] = listFromLine[0:3]
                classLabelVector.append(int(listFromLine[-1]))
                idx += 1
        return returnMat, classLabelVector

def autoNorm(dataSet) :
        minVals = dataSet.min(0)
        maxVals = dataSet.max(0)

        ranges = maxVals - minVals

        normDataSet = zeros(shape(dataSet))
        m = dataSet.shape[0]

        normDataSet = dataSet - tile(minVals,(m,1))
        normDataSet = normDataSet/tile(ranges,(m,1))

        return normDataSet, ranges, minVals

def classify0(inX, dataSet, labels, k) :
        dataSetSize = dataSet.shape[0]
        #calculate the distance between the inx and other traning data
        diffMat = tile(inX, (dataSetSize,1)) - dataSet #titl for array
        sqDiffMat = diffMat**2
        sqDistances = sqDiffMat.sum(axis=1) #calculate the sum
        print "sqDistances=",sqDistances
        distances = sqDistances**0.5
        print "distances=",distances
        sortedDistIndicies = distances.argsort()
        print "sorted=",sortedDistIndicies


#find the k nearest neighbours
        classCount = {}
        for i in range(k) :
                voteIlabel = labels[sortedDistIndicies[i]]
                classCount[voteIlabel] = classCount.get(voteIlabel,0) +1
        sortedClassCount = sorted(classCount.iteritems(), key = operator.itemgetter(1), reverse = True)
        print sortedClassCount
        return sortedClassCount[0][0]

def classfiPerson(filename) :
        resultList = ['not at all', 'in small doses', 'in large doses']

        percentTats = float(raw_input("percentage of time spent playing video games?"))
        ffMiles = float(raw_input("frequent flier miles per year?"))
        iceCream = float(raw_input("ice cream consumed per year?"))

        inArr = array([ffMiles, percentTats, iceCream])

        datingDataMat, datingLabels = file2matrix(filename)

        normMat, ranges, minVals = autoNorm(datingDataMat)
        classifierResult = classify0((inArr-minVals)/ranges, normMat, datingLabels, 3)
        print "result=", resultList[classifierResult-1]


if __name__=='__main__' :
        #group, labels = createDataSet()
        #classify0([0,0],group, labels, 3)
        filename = sys.argv[1]
        returnMat, classLabelVector = file2matrix(filename)
        #print returnMat
        #print classLabelVector


        #normMat, ranges, minVals = autoNorm(returnMat)
        #print ranges
        #print minVals
        #print normMat
        #fig = plt.figure()
        #ax = fig.add_subplot(111)
        #ax.scatter(returnMat[:,1],returnMat[:,2], 15.0*array(classLabelVector), 15.0*array(classLabelVector))
        #plt.show()

        classfiPerson(filename)

（1）调用tile， 复制size个测试变量，与邻居求差值

（2）差值平方，调用sum求和，计算每个的距离

（3）距离按照由小到达排序

（4）遍历前k个距离最近的点，统计分类，找出分类数最大的作为测试变量的分类

四。执行

在终端执行： python KNN.py