python的KNN算法基本实现

数据集获取参考：http://blog.csdn.net/nima1994/article/details/72784545

关键步骤如下：
1.数据处理
通常数据集划分为训练集：测试集=67:33
2.相似度度量
欧氏距离：空间中的实际距离
汉明距离：两个（相同长度）字对应位不同的数量
3.近邻查找
4.预测与测试

代码如下：

import csv
import random
import math
import numpy as np

def euclideanDistance(instance1, instance2, length):
    distance = 0
    for x in range(length):
        distance += pow((instance1[x] - instance2[x]), 2)
    return math.sqrt(distance)
def loadDataset(filename, split, trainingSet=[] , testSet=[]):
    with open(filename) as csvfile:
        lines = csv.reader(csvfile)
        dataset = list(lines)
        for x in range(len(dataset)-1):
            for y in range(length):
                dataset[x][y] = float(dataset[x][y])
            if random.random()<split:
                trainingSet.append(dataset[x])
            else:
                testSet.append(dataset[x])
def getNeighbors(trainingSet, testInstance, length,k):
    neighbors = []
    for x in np.argsort(np.array([euclideanDistance(testInstance, t, length) for t in trainingSet]))[:k]:
        neighbors.append(trainingSet[x])
    return neighbors
def getResponse(neighbors):
    classVotes = {}
    for x in neighbors:
        response = x[-1]
        if response in classVotes:
            classVotes[response] += 1
        else:
            classVotes[response] = 1
    classVotes=sorted(classVotes.items(),key=lambda item:item[1],reverse=True)
    return classVotes[0][0]

def getAccuracy(testSet, predictions):
    correct = 0
    for x in range(len(testSet)):
        if testSet[x][-1]==predictions[x]:
            correct += 1
    return (correct/float(len(testSet))) * 100.0

trainingSet=[]
testSet=[]
predictions=[]
length=4
#初始化训练集和测试集
loadDataset('d:/iris.data', 0.67, trainingSet, testSet)

for x in testSet:
    # 得到近邻
    neighbors = getNeighbors(trainingSet, x, length, 5)
    result = getResponse(neighbors)
    predictions.append(result)
accuracy = getAccuracy(testSet, predictions)
print('Accuracy: ' + repr(accuracy) + '%')

运行结果如：