案例十一 白葡萄酒品质预测

import pandas as pd import numpy as np #Importing The data set red_wine = pd.read_csv("winequality-red.csv",sep=';',engine='python') white_wine = pd.read_csv("winequality-white.csv",sep=';',engine='python') #Checking is there any missing value in data set print("Checking missing values of Red Wine") print(red_wine.isnull().sum()) print("---------------------") print("Checking missing values of White wine") print(white_wine.isnull().sum()) #Printing shape print("Red Wine Shape:-",red_wine.shape) print("------------") print("White Wine Shape:-",white_wine.shape) #Dependent And Independent Variable of Red Wine X_red = red_wine.iloc[: , :-1].values y_red = red_wine.iloc[: , -1].values #Dependent And Independent Variable of White Wine X_white = white_wine.iloc[: , :-1].values y_white = white_wine.iloc[: , -1].values #Feature Scalling on Red Wine from sklearn.preprocessing import StandardScaler sc_Rx = StandardScaler() X_red = sc_Rx.fit_transform(X_red) #Feature Scalling on White Wine from sklearn.preprocessing import StandardScaler sc_Wx = StandardScaler() X_white = sc_Wx.fit_transform(X_white) #Splitting Red wine dataset into training and test set from sklearn.model_selection import train_test_split XR_train,XR_test,yR_train,yR_test = train_test_split(X_red,y_red, random_state=0, test_size=0.25) #Splitting White wine dataset into training and test set from sklearn.model_selection import train_test_split XW_train,XW_test,yW_train,yW_test = train_test_split(X_white,y_white, random_state=0, test_size=0.25) #Training The Red_Wine set from sklearn.svm import SVR regressor = SVR(kernel='rbf') regressor.fit(XR_train,yR_train) #Testing The model Red_predict = regressor.predict(XR_test) errors = abs(Red_predict - yR_test) print('Metrics for SupportVectorRegressor Trained on Red WineData') # Calculate mean absolute percentage error (MAPE) mape = 100 * (errors / yR_test) # Calculate and display accuracy accuracy = 100 - np.mean(mape) print('Accuracy for Red Wine:', round(accuracy, 5), '%.') #Training The White_Wine set from sklearn.svm import SVR regressor = SVR(kernel='rbf') regressor.fit(XW_train,yW_train) #Testing The model White_predict = regressor.predict(XW_test) errors = abs(White_predict - yW_test) print('Metrics for SupportVectorRegressor Trained on White Wine Data') # Calculate mean absolute percentage error (MAPE) mape = 100 * (errors / yW_test) # Calculate and display accuracy accuracy = 100 - np.mean(mape) print('Accuracy for White Wine:', round(accuracy, 5), '%.')