KNN算法和实现

 KNN要用到欧氏距离

 KNN下面的缺点很容易使分类出错（比如下面黑色的点）

 下面是KNN算法的三个例子demo，

第一个例子是根据算法原理实现

import matplotlib.pyplot as plt
import numpy as np
import operator
# 已知分类的数据
x1 = np.array([3,2,1])
y1 = np.array([104,100,81])
x2 = np.array([101,99,98])
y2 = np.array([10,5,2])
scatter1 = plt.scatter(x1,y1,c='r')
scatter2 = plt.scatter(x2,y2,c='b')
# 未知数据
x = np.array([18])
y = np.array([90])
scatter3 = plt.scatter(x,y,c='k')
#画图例
plt.legend(handles=[scatter1,scatter2,scatter3],labels=['labelA','labelB','X'],loc='best')
plt.show()
# 已知分类的数据
x_data = np.array([[3,104],
                   [2,100],
                   [1,81],
                   [101,10],
                   [99,5],
                   [81,2]])
y_data = np.array(['A','A','A','B','B','B'])
x_test = np.array([18,90])
# 计算样本数量
x_data_size = x_data.shape[0]
print(x_data_size)
# 复制x_test
print(np.tile(x_test, (x_data_size,1)))
# 计算x_test与每一个样本的差值
diffMat = np.tile(x_test, (x_data_size,1)) - x_data
diffMat
# 计算差值的平方
sqDiffMat = diffMat**2
sqDiffMat
# 求和
sqDistances = sqDiffMat.sum(axis=1)
sqDistances
# 开方
distances = sqDistances**0.5
print(distances)
# 从小到大排序
sortedDistances = distances.argsort()#返回distances里的数据从小到大的下标数组
print(sortedDistances)
classCount = {}
# 设置k
k = 5
for i in range(k):
    # 获取标签
    votelabel = y_data[sortedDistances[i]]
    # 统计标签数量
    classCount[votelabel] = classCount.get(votelabel,0) + 1#)0表示没有该字典里没有该值时默认为0
classCount
# 根据operator.itemgetter(1)-第1个值对classCount排序，然后再取倒序
sortedClassCount = sorted(classCount.items(),key=operator.itemgetter(1), reverse=True)
print(sortedClassCount)
# 获取数量最多的标签
knnclass = sortedClassCount[0][0]#第一个0表示取第一个键值对('A', 3)，第二个0表示取('A', 3)的‘A’
print(knnclass)

1 import numpy as np#对iris数据集进行训练分类
 2 from sklearn import  datasets
 3 from sklearn.model_selection import  train_test_split
 4 from sklearn.metrics import  classification_report,confusion_matrix#对模型分类结果进行评估的两个模型
 5 import operator#，关于operator的使用
 6 import  random
 7 def knn(x_test, x_data, y_data, k):
 8     x_data_size = x_data.shape[0]  # 计算样本数量
 9     diffMat = np.tile(x_test,(x_data_size,1)) - x_data# 复制x_test，计算x_test与每一个样本的差值
10     sqDiffMat = diffMat**2# # 计算差值的平方
11     sqDistance = sqDiffMat.sum(axis= 1) # 求和
12     distances = sqDistance**0.5 # 开方
13     sortedDistance = distances.argsort()# 从小到大排序
14     classCount = {}
15     for i in range(k):
16         vlabel = y_data[sortedDistance[i]] # 获取标签
17         classCount[vlabel] = classCount.get(vlabel,0)+1# 统计标签数量
18     sortedClassCount = sorted(classCount.items(),key = operator.itemgetter(1), reverse = True) # 根据operator.itemgetter(1)-第1个值对classCount排序，然后再取倒序
19     return  sortedClassCount[0][0]
20 iris = datasets.load_iris()# 载入数据
21 x_train,x_test,y_train,y_test = train_test_split(iris.data, iris.target, test_size=0.3)
22 #打乱数据
23 # data_size = iris.data.shape[0]
24 # index = [i for i in range(data_size)]
25 # random.shuffle(index)
26 # iris.data = iris.data[index]
27 # iris.target = iris.target[index]
28 # test_size = 40#切分数据集
29 # x_train = iris.data[test_size:]
30 # x_test = iris.data[:test_size]
31 # y_train = iris.target[test_size:]
32 # y_test = iris.target[:test_size]
33 prodictions = []
34 for i in range(x_test.shape[0]):
35     prodictions.append(knn(x_test[i],x_train,y_train,5))
36 print(prodictions)
37 print(classification_report(y_test, prodictions))
38 print(confusion_matrix(y_test,prodictions))
39 #关于混淆矩阵可以看这篇博客，

1 # 导入算法包以及数据集
 2 from sklearn import neighbors
 3 from sklearn import datasets
 4 from sklearn.model_selection import train_test_split
 5 from sklearn.metrics import classification_report
 6 import random
 7 # 载入数据
 8 iris = datasets.load_iris()
 9 #print(iris)
10 # 打乱数据切分数据集
11 # x_train,x_test,y_train,y_test = train_test_split(iris.data, iris.target, test_size=0.2) #分割数据0.2为测试数据，0.8为训练数据
12 
13 #打乱数据
14 data_size = iris.data.shape[0]
15 index = [i for i in range(data_size)]
16 random.shuffle(index)
17 iris.data = iris.data[index]
18 iris.target = iris.target[index]
19 
20 #切分数据集
21 test_size = 40
22 x_train = iris.data[test_size:]
23 x_test =  iris.data[:test_size]
24 y_train = iris.target[test_size:]
25 y_test = iris.target[:test_size]
26 
27 # 构建模型
28 model = neighbors.KNeighborsClassifier(n_neighbors=3)
29 model.fit(x_train, y_train)
30 prediction = model.predict(x_test)
31 print(prediction)
32 print(classification_report(y_test, prediction))

这三个代码第一个，第二个是根据底层原理实现knn算法，第三个则是调用库函数处理数据。

 下面一个代码是利用第三个代码中用到的库实现第一个代码功能，可以发现使用系统提供的库，简单许多

1 from sklearn import  neighbors
 2 from sklearn.model_selection import train_test_split
 3 from sklearn.metrics import classification_report
 4 import numpy as np
 5 x_data = np.array([[3,104],
 6                    [2,100],
 7                    [1,81],
 8                    [101,10],
 9                    [99,5],
10                    [81,2]])
11 y_data = np.array(['A','A','A','B','B','B'])
12 x_test1 = np.array([[18,90]])
13 x_train, x_test, y_train,y_test = train_test_split(x_data, y_data,test_size= 0.3)
14 model = neighbors.KNeighborsClassifier(n_neighbors=3)
15 model.fit(x_train, y_train)
16 print(x_test1)
17 prediction = model.predict(x_test1)
18 print(prediction)

对字典进行排序：dic = sorted(dic.items(),key = operator.itemgetter(1),reverse=True)

作者：你的雷哥

本文版权归作者所有，欢迎转载，但未经作者同意必须在文章页面给出原文连接，否则保留追究法律责任的权利。