k近邻算法分类

k近邻算法分类
X= l[[‘PM2.5’,‘PM10’,‘CO’,‘NO2’,‘O3’,‘SO2’]] #选取列名为x1,x2,x3的列作为X
y= l[[‘AQI’]] #选取列名为y的作为Y.
from sklearn.neighbors import KNeighborsRegressor
from sklearn.metrics import r2_score,mean_absolute_error
import numpy as np
X_train=X.values
y_train=y.values
train = pd.read_csv("./2019dongsi.csv")
train.head()
Unnamed: 0 date hour station PM2.5 PM10 CO AQI NO2 O3 SO2
0 0 20190101 0 东四 68.0 82.0 1.6 91.0 74.0 7.0 16.0
1 1 20190101 1 东四 57.0 81.0 1.4 78.0 74.0 7.0 12.0
2 2 20190101 2 东四 47.0 74.0 1.4 65.0 75.0 7.0 9.0
3 3 20190101 3 东四 46.0 75.0 1.2 64.0 73.0 7.0 8.0
4 4 20190101 4 东四 46.0 70.0 1.2 64.0 72.0 7.0 7.0
X1= train[[‘PM2.5’,‘PM10’,‘CO’,‘NO2’,‘O3’,‘SO2’]] #选取列名为x1,x2,x3的列作为X
y1= train[[‘AQI’]]
X_test=X1.values
y_test=y1.values
X_test.shape
(6853, 6)
len(y_test)
6853
y_train1=y_train[0:len(y_test)] plt.rcParams[“font.family”]=“SimHei” plt.plot(np.linspace(0,6853,6853),y_train1,".-r",lw=1.5,mec=“k”,ms=8) plt.plot(np.linspace(0,6853,100),y_test,".-g",lw=1.5,mec=“k”,ms=8) plt.legend([“训练集”,“测试集”],loc=0,numpoints=3,fontsize=12) plt.xlabel(“样本个数”,fontsize=12) plt.ylabel(“y值”,fontsize=12)

X= l[[‘PM2.5’,‘PM10’,‘CO’,‘NO2’,‘O3’,‘SO2’]] #选取列名为x1,x2,x3的列作为X
y= l[[‘AQI’]]
X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.2,random_state=0)
knn_reg=KNeighborsRegressor(n_neighbors=15,
weights=“distance”,
metric=“minkowski”,
p=2)
knn_reg.fit(X_train,y_train)
KNeighborsRegressor(algorithm=‘auto’, leaf_size=30, metric=‘minkowski’,
metric_params=None, n_jobs=1, n_neighbors=15, p=2,
weights=‘distance’)
knn_reg.metric,knn_reg.algorithm,knn_reg.effective_metric_
(‘minkowski’, ‘auto’, ‘euclidean’)
y_pred=knn_reg.predict(X_test)
score=knn_reg.score(X_test,y_test)
score
0.9972043343639387
plt.plot(np.linspace(0,6853,6853),y_test,".-r",lw=1.5,mec=“k”,ms=8) plt.plot(np.linspace(0,6853,100),y_pred,".-3",lw=1.5,mec=“k”,ms=8) plt.legend([“训练集的真实值”,“测试集的预测值”],loc=0,numpoints=3,fontsize=12) plt.xlabel(“样本个数”,fontsize=12) plt.ylabel(“y值”,fontsize=12)

knn_reg.score(X_test,y_test)
0.9972043343639387
r2_score(y_test,y_pred)
0.9972043343639387
mean_squared_error(y_test,y_pred)
15.37593152752547
from sklearn.model_selection import KFold
kf = KFold(n_splits=3)
for train,test in kf.split(X):
print(f"训练集是:{train},测试集是:{test}")