KNN算法和sklearn中的KNN算法

# -*- coding: utf-8 -*-
import numpy as np
from math import sqrt
from collections import Counter

class KNN_Classifier:
    def __init__(self,k):
        """构造函数"""
        #给定k值
        self.k=k
        self.X_train=None
        self.y_train=None
    def fit(self,X_train,y_train):
        """训练模型"""
       #knn的训练模型实际上就是把训练数据集进行保存
        assert self.k<=X_train.shape[0],\
               "k的值必须小于等于X_train的第二个维度"
        self._X_train=X_train
        self._y_train=y_train
        return self
    def predicty(self,x):
        """对矩阵中的元素进行预测"""
        assert self._X_train is not None and self._y_train is not None,\
               "X_train and y_train must be fit"
        assert self._X_train.shape[1]==x.shape[1],\
               "X_train和x的维度应该相等"
        y_predict=[self._predict(m) for m in x]
        return np.array(y_predict)
    def _predict(self,m):
        """预测每个点的函数"""
        #求距离
        distances=np.array(np.array([sqrt(np.sum((x-m)**2)) for x in self._X_train]))
        #得到值从小到大后排序的索引值
        top_y=distances.argsort()
        #取出最近的k个值
        nearest=self._y_train[top_y[:self.k]]
        #得到预测值
        predict_y=Counter(nearest).most_common(1)[0][0]
        return predict_y

文章源自随机的未知-https://sjdwz.com/11100.html

import numpy as np
from sklearn.neighbors import KNeighborsClassifier
import matplotlib.pyplot as plt

raw_data_X = [
    [3.393533211, 2.331273381],
    [3.110073483, 1.781539638],
    [1.343808831, 3.368360954],
    [3.582294042, 4.679179110],
    [2.280362439, 2.866990263],
    [7.423436942, 4.696522875],
    [5.745051997, 3.533989803],
    [9.172168622, 2.511101045],
    [7.792783481, 3.424088941],
    [7.939820817, 0.791637231]
]  # 两列分别代表肿瘤的长和宽 X是大写
raw_data_y = [0, 0, 0, 0, 0, 1, 1, 1, 1, 1] # 分为0,1两个类别
#x训练集
X_train=np.array(raw_data_X)
y_train=np.array(raw_data_y)

#看一下原始数据点的分布
plt.figure(1)
plt.scatter(X_train[y_train==0,0],X_train[y_train==0,1],color='red',marker='o',label='y==0')
plt.scatter(X_train[y_train==1,0],X_train[y_train==1,1],color='blue',marker='x',label='y==1')
plt.legend()
plt.show()

X_new = np.array([8.093623049, 3.365732334])#增加一个新的数据
#用散点图直观观察一下新增数据之后新的数据的分布
plt.figure(2)
plt.scatter(X_train[y_train==0,0],X_train[y_train==0,1],color='red',marker='o',label='y==0')
plt.scatter(X_train[y_train==1,0],X_train[y_train==1,1],color='blue',marker='x',label='y==1')
plt.legend()
plt.scatter(X_new[0],X_new[1],marker='+',color='green',s=100)
plt.show()

"""sklearn中的KNN"""

#创建knn类

knn=KNeighborsClassifier()
#训练knn
knn.fit(X_train,y_train)
#将X_new变成两列的
X_new=X_new.reshape(-1,2)
#输出类别
print(knn.predict(X_new))

文章源自随机的未知-https://sjdwz.com/11100.html

import numpy as np
from sklearn import datasets
from sklearn.model_selection import train_test_split
from sklearn.model_selection import GridSearchCV
from sklearn.preprocessing import StandardScaler
from sklearn.neighbors import KNeighborsClassifier

#加载 sklearn 中的 datasets 数据中的digits（手写字体数据）

digits=datasets.load_digits()
x=digits.data
y=digits.target

#使用 train_test_split 进行模型建立和测试的分离,加zhiqian是因为没有进行
#归一化处理

X_train_zhiqian,X_test_zhiqian,y_train,y_test=train_test_split(x,y,test_size=0.2)

#使用 StandardScaler进行归一化处理
"""为什么使用X_train_zhiqian进行归一化处理的均值和方差进行对X_test的归一化呢？
是因为要服从现实情况，如果只有一个数据进行预测，就没有办法得出均值和方差
"""

s=StandardScaler()
s.fit(X_train_zhiqian)
X_train=s.transform(X_train_zhiqian)
X_test=s.transform(X_test_zhiqian)

knn_clf=KNeighborsClassifier()
#使用网格搜索的方法进行对超参数的选取
#超参数是指在算法进行时对不确定的参数进行调整以使得算法的预测准确率最高
params=[
{"weights":["uniform"],
 "n_neighbors":[k for k in range(1,11)]
},
{"weights":["distance"],
 "n_neighbors":[k for k in range(1,11)],
 "p":[k for k in range(1,6)]
}
]

#n_jobs 这个参数是使用计算机的几个核，赋值为-1表示全都使用

gridsearch=GridSearchCV(knn_clf,params,n_jobs=-1)
gridsearch.fit(X_train,y_train)
print("best_score",gridsearch.best_score_)
print("best_params",gridsearch.best_params_)

#将最好的赋值给knn_clf
knn_clf=gridsearch.best_estimator_

print(knn_clf.score(X_test,y_test))

文章源自随机的未知-https://sjdwz.com/11100.html