python3 learn machine learning api
Use two k-nearest neighbor regression models, average k-nearest neighbor regression and distance-weighted k-nearest neighbor regression, for prediction.
git: https://github.com/linyi0604/MachineLearning
Code:
1 from sklearn.datasets import load_boston 2 from sklearn.cross_validation import train_test_split 3 from sklearn.preprocessing import StandardScaler 4 from sklearn.neighbors import KNeighborsRegressor 5 from sklearn.metrics import r2_score, mean_squared_error, mean_absolute_error 6 import numpy as np 7 8 # 1 准备数据 9 # 读取波士顿地区房价信息 10 boston = load_boston() 11 #View the data description 12 # print(boston.DESCR) # A total of 506 pieces of housing price information in the Boston area, each with 13 numerical feature descriptions and target house prices 13 #Check the difference between the data 14 # print("Maximum house price:", np .max(boston.target)) # 50 15 # print("Minimum house price:",np.min(boston.target)) # 5 16 # print("Average house price:", np.mean(boston.target)) # 22.532806324110677 17 18 x = boston.data 19 y = boston.target 20 21 # 2 split training data and test data 22 # randomly sample 25% as test 75% as training 23 x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.25, random_state=33 ) 24 25 26 # 3 Normalize training data and test data 27 ss_x = StandardScaler() 28 x_train = ss_x.fit_transform(x_train) 29 x_test = ss_x.transform(x_test) 30 31 ss_y = StandardScaler() 32 y_train = ss_y. fit_transform(y_train.reshape(-1, 1 )) 33 y_test = ss_y.transform(y_test.reshape(-1, 1 )) 34 35 # 4 Two kinds of k-nearest neighbor regression line learning and prediction 36 # Initialize k-nearest neighbor regression model using Average regression for prediction 37 uni_knr = KNeighborsRegressor(weights= " uniform " ) 38 #training 39 uni_knr.fit (x_train, y_train) 40 #predicting and saving the prediction results 41 uni_knr_y_predict = uni_knr.predict(x_test) 42 43 # multi-initialization k-nearest neighbor regression model using distance weighted regression 44 dis_knr = KNeighborsRegressor(weights= " distance " ) 45 #training 46 dis_knr.fit (x_train, y_train) 47 #predicting and saving prediction results 48 dis_knr_y_predict = dis_knr.predict (x_test) 49 50 # 5 model evaluation 51 #average k-nearest neighbor regression model evaluation 52 print ( " The default evaluation value of the average k-nearest neighbor regression is: " , uni_knr.score(x_test, y_test)) 53 print ( " The R_squared value of the average k-nearest neighbor regression is: " , r2_score(y_test, uni_knr_y_predict)) 54 print ( "The average k The mean squared error of nearest neighbor regression is: " , mean_squared_error(ss_y.inverse_transform(y_test), 55 ss_y.inverse_transform(uni_knr_y_predict))) 56 print ( " The mean absolute error of mean k nearest neighbor regression is: " , mean_absolute_error(ss_y.inverse_transform( y_test), 57 ss_y.inverse_transform(uni_knr_y_predict))) 58 # Distance-weighted k-nearest neighbor regression model evaluation 59 print ( " The default evaluation value of distance-weighted k-nearest neighbor regression is: " , dis_knr.score(x_test, y_test)) 60 print ( " The R_squared value of distance-weighted k-nearest neighbor regression is: " , r2_score(y_test, dis_knr_y_predict)) 61 print ( " The mean squared error of the distance-weighted k-nearest neighbor regression is: " , mean_squared_error(ss_y.inverse_transform(y_test), 62 ss_y.inverse_transform(dis_knr_y_predict))) 63 print ( "The distance-weighted k-nearest neighbor The mean absolute error of the regression is: " , mean_absolute_error(ss_y.inverse_transform(y_test), 64 ss_y.inverse_transform(dis_knr_y_predict))) 65 66 ''' 67 Mean k-nearest neighbor regression default evaluation value: 0.6903454564606561 68 Mean k-nearest neighbor regression R_squared value: 0.6903454564606561 69 Mean squared error of mean k-nearest neighbor regression: 24.0110141 The mean absolute error of k-nearest neighbor regression is: 2.9680314960629928 71 The default evaluation value of distance-weighted k-nearest neighbor regression is: 0.7197589970156353 72 The R_squared value of distance-weighted k-nearest neighbor regression is: 0.7197589970156353 73 The mean squared error of distance-weighted k-nearest neighbor regression is: 21.73025604 72 The mean absolute error of weighted k-nearest neighbor regression is: 2.8050568785108005 75 '''