ML_KNN

KNN

概要：

K邻近算法：将输入样本归为最近的K个样本中所属类别最多的一个类

三个基本要素：K的选择，距离的定义，分类决策的规则

K邻近算法的实现（动手构造才能更加理解算法）：kd树

 

 用kd树的最邻近搜索：

实现knn_classify:

#%%计算原理
from sklearn import datasets
from collections import Counter  # 为了做投票
from sklearn.model_selection import train_test_split
import numpy as np

# 导入iris数据
iris = datasets.load_iris()
X = iris.data
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=20,test_size=0.3)
# =============================================================================
# #help(train_test_split)
# #random_state 相当于随机数种子,这里的random_state就是为了保证程序每次运行都分割一样的训练集和测试集。否则，同样的算法模型在不同的训练集和测试集上的效果不一样。
def dis(instance1,instance2):
    """
    欧氏距离
    """
    distan=np.sqrt(sum((instance1-instance2))**2)
    return distan

def class_knn(X,y,instance,k):
    """
    k不同结果会有变化
    将实例instance和x_train中的实例计算距离，找到k个最小距离对应的样本归属的最多的类别
    """
    distance=[dis(x,instance) for x in X]
    neighbors=np.argsort(distance)[:k]
    count=Counter(y[neighbors])
    return count.most_common()[0][0]
# =============================================================================
#collections.Counter:Help on class Counter 
    #most_common() list all unique
# #np.argsort(-x) #按降序排列
    # x[np.argsort(x)] #通过索引值排序后的数组
# =============================================================================
predictions=[class_knn(X_train,y_train,data,3) for data in X_test]
np.vstack((y_test,predictions))  #合并比较查看
correct=np.count_nonzero((predictions==y_test)==True)
print("Accurancy is: %.3f"%(correct/len(X_test)))

#%%调包
import numpy as np
from sklearn import datasets
from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier
iris=datasets.load_iris()
x=iris.data
y=iris.target
x_train,x_test,y_train,y_test=train_test_split(x,y,test_size=0.3)
#定义模型
knn=KNeighborsClassifier(n_neighbors=3)  #k=3
knn.fit(x_train,y_train)  #fit拟合
knn.predict(x_test)
knn.classes_
knn.get_params
# 计算准确率
from sklearn.metrics import accuracy_score
correct = np.count_nonzero((knn.predict(X_test)==y_test)==True)
#accuracy_score(y_test, clf.predict(X_test))
print ("Accuracy is: %.3f" %(correct/len(X_test)))

KNN——回归中的表现_learing