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([,,])
y1 = np.array([,,])
x2 = np.array([,,])
y2 = np.array([,,])
scatter1 = plt.scatter(x1,y1,c='r')
scatter2 = plt.scatter(x2,y2,c='b')
# 未知数据
x = np.array([])
y = np.array([])
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([[,],
                   [,],
                   [,],
                   [,],
                   [,],
                   [,]])
y_data = np.array(['A','A','A','B','B','B'])
x_test = np.array([,])
# 计算样本数量
x_data_size = x_data.shape[]
print(x_data_size)
# 复制x_test
print(np.tile(x_test, (x_data_size,)))
# 计算x_test与每一个样本的差值
diffMat = np.tile(x_test, (x_data_size,)) - x_data
diffMat
# 计算差值的平方
sqDiffMat = diffMat**
sqDiffMat
# 求和
sqDistances = sqDiffMat.sum(axis=)
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 =
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表示没有该字典里没有该值时默认为0
classCount
# 根据operator.itemgetter()-第1个值对classCount排序，然后再取倒序
sortedClassCount = sorted(classCount.items(),key=operator.itemgetter(), reverse=True)
print(sortedClassCount)
# 获取数量最多的标签
knnclass = sortedClassCount[][]#第一个0表示取第一个键值对('A', )，第二个0表示取('A', )的‘A’
print(knnclass)

 1 import numpy as np#对iris数据集进行训练分类
 from sklearn import  datasets
 from sklearn.model_selection import  train_test_split
 from sklearn.metrics import  classification_report,confusion_matrix#对模型分类结果进行评估的两个模型
 import operator#https://blog.csdn.net/u010339879/article/details/98304292，关于operator的使用
 import  random
 def knn(x_test, x_data, y_data, k):
     x_data_size = x_data.shape[]  # 计算样本数量
     diffMat = np.tile(x_test,(x_data_size,)) - x_data# 复制x_test，计算x_test与每一个样本的差值
     sqDiffMat = diffMat**# # 计算差值的平方
     sqDistance = sqDiffMat.sum(axis= ) # 求和
     distances = sqDistance**0.5 # 开方
     sortedDistance = distances.argsort()# 从小到大排序
     classCount = {}
     for i in range(k):
         vlabel = y_data[sortedDistance[i]] # 获取标签
         classCount[vlabel] = classCount.get(vlabel,)+# 统计标签数量
     sortedClassCount = sorted(classCount.items(),key = operator.itemgetter(), reverse = True) # 根据operator.itemgetter()-第1个值对classCount排序，然后再取倒序
     return  sortedClassCount[][]
 iris = datasets.load_iris()# 载入数据
 x_train,x_test,y_train,y_test = train_test_split(iris.data, iris.target, test_size=0.3)
 #打乱数据
 # data_size = iris.data.shape[]
 # index = [i for i in range(data_size)]
 # random.shuffle(index)
 # iris.data = iris.data[index]
 # iris.target = iris.target[index]
 # test_size = #切分数据集
 # x_train = iris.data[test_size:]
 # x_test = iris.data[:test_size]
 # y_train = iris.target[test_size:]
 # y_test = iris.target[:test_size]
 prodictions = []
 for i in range(x_test.shape[]):
     prodictions.append(knn(x_test[i],x_train,y_train,))
 print(prodictions)
 print(classification_report(y_test, prodictions))
 print(confusion_matrix(y_test,prodictions))
 #关于混淆矩阵可以看这篇博客，#https://www.cnblogs.com/missidiot/p/9450662.html

 # 导入算法包以及数据集
 from sklearn import neighbors
 from sklearn import datasets
 from sklearn.model_selection import train_test_split
 from sklearn.metrics import classification_report
 import random
 # 载入数据
 iris = datasets.load_iris()
 #print(iris)
 # 打乱数据切分数据集
 # 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，.8为训练数据

 #打乱数据
 data_size = iris.data.shape[]
 index = [i for i in range(data_size)]
 random.shuffle(index)
 iris.data = iris.data[index]
 iris.target = iris.target[index]

 #切分数据集
 test_size =
 x_train = iris.data[test_size:]
 x_test =  iris.data[:test_size]
 y_train = iris.target[test_size:]
 y_test = iris.target[:test_size]

 # 构建模型
 model = neighbors.KNeighborsClassifier(n_neighbors=)
 model.fit(x_train, y_train)
 prediction = model.predict(x_test)
 print(prediction)
 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

 from sklearn import  neighbors
 from sklearn.model_selection import train_test_split
 from sklearn.metrics import classification_report
 import numpy as np
 x_data = np.array([[,],
                    [,],
                    [,],
                    [,],
                    [,],
                    [,]])
 y_data = np.array(['A','A','A','B','B','B'])
 x_test1 = np.array([[,]])
 x_train, x_test, y_train,y_test = train_test_split(x_data, y_data,test_size= 0.3)
 model = neighbors.KNeighborsClassifier(n_neighbors=)
 model.fit(x_train, y_train)
 print(x_test1)
 prediction = model.predict(x_test1)
 print(prediction)