-
CRIM: 城镇人均犯罪率(%)
-
ZN: 住宅用地所占比例(%)
-
INDUS: 城镇中非住宅用地所占比例(%)
-
CHAS: 0-1分类变量,是否靠近Charles River,靠近1,否则0
-
NOX: 一氧化氮指数
-
RM: 每栋住宅的房间数
-
AGE: 1940 年以前建成的自住单位的比例(%)
-
DIS: 距离 5 个波士顿的就业中心的加权距离
-
RAD: 距离高速公路的便利指数
-
TAX: 每一万美元的不动产税率(%)
-
PTRATIO: 城镇中的教师学生比例(%)
-
B: 关于黑人比例的一个参数(%)
-
LSTAT: 地区中有多少房东属于低收入人群(%)
-
MEDV: 自住房屋房价中位数(也就是均价,单位千美元)
代码展示 import numpy as np import matplotlib.pyplot as plt import pandas as pd from sklearn.metrics import mean_squared_error from sklearn.neighbors import KNeighborsRegressor from sklearn.model_selection import train_test_split data = pd.read_csv(r"C:\Users\Lenovo\Desktop\boston_train.csv") x_data =data.iloc[:,:-1] # 导入所有特征变量 y_data = data.iloc[:,-1] # 导入目标值(房价) # print(x_data) # print(y_data)
#选择与房价最相关的三个变量,不 from sklearn.feature_selection import SelectKBest,f_regression selector=SelectKBest(f_regression,k=3) x_new=selector.fit_transform(x_data,y_data)
a=selector.get_support(indices=True).tolist() print(a)#三种最相关的属性是indus rm lstat
x_train1,x_test1,y_train1,y_test1=train_test_split(x_new,y_data,test_size=0.2) print(x_train1.shape)#划分训练集测试集 #归一化 from sklearn.preprocessing import StandardScaler scaler=StandardScaler()#用于计算训练数据的均值和方差 scaler.fit(x_train1) x_train1_std=scaler.transform(x_train1) x_test1_std=scaler.transform(x_test1)
knn_r=KNeighborsRegressor(n_neighbors=5) knn_r.fit(x_train1_std,y_train1)#训练模型 accuracy1=knn_r.score(x_test1_std,y_test1)#模型精确度 print(accuracy1) 准确度: y_pre=knn_r.predict(x_test1_std)#根据测试集预测房价
list_1=[i for i in range(0,y_test1.shape[0])]#x轴样本标号
plt.plot(list_1,y_test1,color="green",linestyle="dashed",markerfacecolor='black',markersize=10, label='True') plt.plot(list_1,y_pre,color="red",linestyle="dashed",markerfacecolor='blue',markersize=10, label='Knn') plt.xlabel('sample') plt.ylabel('price') plt.legend() plt.show()
right=[] for i in range(1,30): knn_r=KNeighborsRegressor(n_neighbors=i) knn_r.fit(x_train1_std,y_train1) accuracy1=knn_r.score(x_test1_std,y_test1) right.append(accuracy1) plt.plot(range(1,30),right,color="black",linestyle="dashed",markerfacecolor='red',markersize=10) plt.xlabel('K Value') plt.ylabel('Accuracy') plt.show()#绘不同k值制准确度