KNN 最近邻

K：距离预测点最近的K个值，哪边多算哪边

监督学习

使用分类算法，且数据比较大

导入包

import pandas as pd
from sklearn import neighbors
import numpy as np
%matplotlib inline  
import seaborn

training_data = pd.DataFrame()
training_data['test_1'] = [0.3051,0.4949,0.6974,0.3769,0.2231,0.341,0.4436,0.5897,0.6308,0.5]
training_data['test_2'] = [0.5846,0.2654,0.2615,0.4538,0.4615,0.8308,0.4962,0.3269,0.5346,0.6731]
training_data['outcome'] = ['win','win','win','win','win','loss','loss','loss','loss','loss']

training_data.head()

绘图

seaborn.lmplot('test_1', 'test_2', data=training_data, fit_reg=False,hue="outcome", scatter_kws={"marker": "D","s": 100})

转化数据为numpy

test12 = training_data[['test_1', 'test_2']]
test12

X = test12.values
y = np.array(training_data['outcome'])

模型训练

#模型拟合
clf = neighbors.KNeighborsClassifier(3, weights = 'uniform')
trained_model = clf.fit(X, y)

评价模型

trained_model.score(X, y)

预测

# 创建一个新观测，为 .4 和 .6
x_test = np.array([[.4,.6]])

# 将学习者应用于新的未分类的观测。
trained_model.predict(x_test)

# array(['loss'], dtype=object) 
trained_model.predict_proba(x_test)

# array([[ 0.66666667,  0.33333333]]) 

K值取值

from matplotlib import pyplot as plt
import numpy as np

score = []
krange = range(1,10)

for i in krange:
    clf = KNeighborsClassifier(n_neighbors=i)
    clf = clf.fit(X,y)
    score.append(clf.score(X,y))
    

plt.xticks(np.array(krange))
plt.plot(krange,score)
plt.figure(figsize=(12,9))
plt.show()

参考链接

https://dsai-notes.apachecn.org/#/14