kNN分类算法

一、概述

    kNN算法采用测量不同特征值之间的距离方法进行分类。对未知类别属性的数据集中的每个点执行如下操作：

    （1）计算已知类别数据集中的点与当前点之间的距离；

    （2）按照距离递增次序排序；

    （3）选取与当前点距离最小的k个点；

    （4）确定前K个点所在类别的出现频率；

    （5）返回前k个点出现频率最高的类别作为当前点的预测分类。

二、代码实现

    1.基于scikit-learn包实现

import numpy as np
from sklearn import neighbors

def split_data(data, test_size):
    data_num = data.shape[0]
    train_ind = list(range(data_num))
    test_ind = []
    test_num = int(data_num * test_size)
    for i in range(test_num):
        rand_ind = np.random.randint(0, len(train_ind))
        test_ind.append(rand_ind)
        del train_ind[rand_ind]
    train_data = data[train_ind]
    test_data = data[test_ind]
    return train_data, test_data

# load the data and divide the data
mydata = np.loadtxt(open("iris.txt","rb"), delimiter = ",", skiprows = 0)

train_data, test_data = split_data(mydata, 0.3)
n = mydata.shape[1]
test_label = test_data[:, n-1]
test_data = test_data[:,0:n-1]

train_label = train_data[:,n-1]
train_data = train_data[:, 0:n-1]

# get the KNN classifier
knn = neighbors.KNeighborsClassifier()

knn.fit(train_data, train_label)
print(knn.predict(test_data))

运行结果如下：

2、python代码逐步实现

import numpy as np
import operator

def split_data(data, test_size):
    data_num = data.shape[0]
    train_ind = list(range(data_num))
    test_ind = []
    test_num = int(data_num * test_size)
    for i in range(test_num):
        rand_ind = np.random.randint(0, len(train_ind))
        test_ind.append(rand_ind)
        del train_ind[rand_ind]

    train_data = data[train_ind]
    test_data = data[test_ind]
    return train_data, test_data


def createDataSet():
    group = np.array([[1.0, 1.1], [1.0, 1.0], [0, 0], [0, 0.1]])
    # labels = ['A', 'A', 'B', 'B']
    labels = np.array([1, 1, 2, 2])
    # print (group)
    # print (labels)
    return group, labels


def classify0(inX, dataSet, labels, k):
    dataSetSize = dataSet.shape[0]
    diffMat = np.tile(inX, (dataSetSize, 1)) - dataSet
    sqDiffMat = diffMat ** 2
    sqDistances = sqDiffMat.sum(axis=1)
    distances = sqDistances ** 0.5
    sortedDistIndicies = distances.argsort()
    classCount = {}
    for i in range(k):
        voteIlabel = labels[sortedDistIndicies[i]]
        classCount[voteIlabel] = classCount.get(voteIlabel, 0) + 1
    sortedClassCount = sorted(classCount.items(), key = operator.itemgetter(1), reverse = True)
    return sortedClassCount[0][0]


def classify1(inX, dataSet, labels, k):
    result_ind = []
    inX_size = inX.shape[0]
    dataSetSize = dataSet.shape[0]
    for i in range(inX_size):
        diffMat = np.tile(inX[i,:], (dataSetSize, 1)) - dataSet
        sqDiffMat = diffMat ** 2
        sqDistances = sqDiffMat.sum(axis=1)
        distances = sqDistances ** 0.5
        sortedDistIndicies = distances.argsort()
        classCount = {}
        for j in range(k):
            voteIlabel = labels[sortedDistIndicies[j]]
            classCount[voteIlabel] = classCount.get(voteIlabel, 0) + 1
        sortedClassCount = sorted(classCount.items(), key=operator.itemgetter(1), reverse=True)
        ind = sortedClassCount[0][0]
        result_ind.append(ind)
    return result_ind


# load the data and divide the data
mydata = np.loadtxt(open("iris.txt","rb"), delimiter = ",", skiprows = 0)

train_data, test_data = split_data(mydata, 0.3)
n = mydata.shape[1]
test_label = test_data[:, n-1]
test_data = test_data[:,0:n-1]

train_label = train_data[:,n-1]
train_data = train_data[:, 0:n-1]

# test code -- classify 0
result_ind = []
for i in range(len(test_data)):
    ind = classify0(test_data[i,:], train_data, train_label,7)
    result_ind.append(ind)

print(result_ind)
#
#
# # test code -- classify 1
# result_index = classify1(test_data, train_data, train_label, 3)
# print(result_index)

运行结果如下：