雪梨k近邻作业

ML作业.md

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.externals import joblib

iris = joblib.load('later.pkl')
data = pd.DataFrame(iris,index = range(0,150),columns=['a','b','c','d','e'])
X = data.loc[:,'a':'d']
y = data['e']

from sklearn.model_selection import cross_val_score
from sklearn.neighbors import KNeighborsClassifier
knn_clf = KNeighborsClassifier()

mean=[]
std =[] 

for k in {1,4,7,10,13,16}:
    knn_clf = KNeighborsClassifier(weights = "distance",n_neighbors = k,p = 2 )# weights = "uniform"
    scores = cross_val_score(knn_clf,X,y,cv=5)
    #print(scores)
    score_m = np.mean(scores)
    mean.append(score_m)
    score_s = np.std(scores)
    std.append(score_s)

plt.rcParams["axes.grid"] = True
plt.rcParams["grid.linestyle"] = (5,9)
fig = plt.figure(figsize=(8,6),dpi=100)
ax = plt.axes()
l = [1,4,7,10,13,16]

l1 = plt.plot(l, mean, color = '#3a3afa', linewidth=2.0, linestyle='-') 

dy = np.array(mean)-np.array(std)
plt.errorbar(l, mean, yerr=std, color = '#3a3afa')
plt.show()

加载数据函数

import numpy as np 
import matplotlib.pyplot as plt 
from sklearn import neighbors,datasets
from sklearn.model_selection import train_test_split

def load_classification_data():
    digits=datasets.load_digits() 
    X_train=digits.data 
    y_train=digits.target 
    return train_test_split(X_train, y_train,test_size=0.25,random_state=0,stratify=y_train)

def create_regression_data(n): # create_regression_data函数是在sin(X)基础上添加噪声生成的。
    X =5 * np.random.rand(n, 1) 
    y = np. sin(X). ravel() 
    y[::5] += 1 * (0.5 -np.random.rand(int(n/5))),  
    return train_test_split(X, y,test_size=0.25,random_state=0)

scikit-learn中提供了一个KNeighborsClassifier类来实现k近邻法分类模型，其原型为：
sklearn.neighbors.KNeighborsClassifier(n_neighbors=5, weights='uniform', algorithm='auto', leaf_size=30, p=2, metric='minkowski', metric_params=None, n_j obs=1, **kwargs)

参数：

1. n_neighbors:int，指定k值
2. weights:一字符串或者可调用对象，指定投票权重类型。即这些邻居投票权可以为相同或者不同。
   • 'uniform'：本节点的所有邻居节点的投票权重都相等。
   • 'distance'：本节点的所有邻居节点的投票权重与距离成反比。即越近的节点，其投票权重越大。
   • [callable]：一个可调用对象。它传入距离的数组，返回同样形状的权重数组。
3. algorithm:一字符串，指定计算最近邻的算法，可以为如下：
   • 'ball_tree'：使用BallTree算法。
   • 'kd_tree:使用KDTree算法。
   • 'brute'：使用暴力搜索法。
   • 'auto'：自动决定最合适的算法。
4. leaf _size :指定BallTree或者KDTree叶节点规模。它影响树的构建和查询速度
5. metric:str，指定距离度量，default'minkowski'距离
6. p:int指定在'Minkowski'度量上的指数。如果p=1，对应曼哈顿距离；如果'p=2'对应欧拉距离
7. n_jobs:并行性。default：-1表示派发任务到所有计算机CPU上

Methods:

1. fit(X,y):训练模型
2. predict(X)使用model来预测，返回带预测Sample的标记
3. score(X,y)返回在(X,y)上预测的准确率accuracy
4. predict_proba(X)返回Sample为每种标记的概率
5. kneighbors([X,n_neighbors,return_distance])返回Sample point 的k近邻点。如果return_distance=True，同时返回到这些紧邻点的距离。
6. kneighbors_graph([X,n_neighbors,mode])返回样本点的连接图

def test_KNeighborsClassifier(*data): #首先使用KNeighborsClassifier,给出测试函数：
    X_train,X_test,y_train,y_test=data
    clf = neighbors.KNeighborsClassifier()
    clf.fit(X_train,y_train)
    print("Train Score:%f"%clf.score(X_train,y_train))
    print("Testing Score:%f"%clf.score(X_test,y_test))

X_train,X_test,y_train,y_test=load_classification_data()
test_KNeighborsClassifier(X_train,X_test,y_train,y_test) #Train Score:0.991091 #Testing Score:0.980000

def test_KNeighborsClassifier_k_w(*data):
    X_train,X_test,y_train,y_test=data
    Ks=np.linspace(1,y_train.size,num=100,endpoint=False,dtype='int')
    weights=['uniform','distance']
    
    fig = plt.figure()
    ax = fig.add_subplot(1,1,1)
    for weight in weights:
        training_scores=[]
        testing_scores=[]
        for K in Ks:
            clf=neighbors.KNeighborsClassifier(weights=weight,n_neighbors=K)
            clf.fit(X_train,y_train)
            testing_scores.append(clf.score(X_test,y_test))
            training_scores.append(clf.score(X_train,y_train))
        
        ax.plot(Ks,testing_scores,label="testing score:weight=%s"%weight)
        ax.plot(Ks,training_scores,label="training score:weight=%s"%weight)
    ax.legend(loc='best')
    ax.set_xlabel('K')
    ax.set_ylabel("score")
    ax.set_ylim(0,1.05)
    ax.set_title("KNeighborsClassifier")
    plt.show()

test_KNeighborsClassifier_k_w(X_train,X_test,y_train,y_test)
    ````