本文共 4027 字，大约阅读时间需要 13 分钟。

单个特征的回归问题：

from sklearn import linear_modelimport numpy as npimport pandas as pdimport matplotlib.pyplot as pltfrom sklearn import datasets,linear_model#从csv文件读取数据的函数 ，这里直接用了数据，没有读取def get_data(file_name): data = pd.read_csv(file_name) X_parameter = [] Y_parameter = [] for single_square_feet ,single_price_value in zip(data['size'],data['price']):       X_parameter.append([float(single_square_feet)])       Y_parameter.append(float(single_price_value)) return X_parameter,Y_parameter'''    构建一个 一size,price为列名的csv数据    数据转化为csv格式的方法    在dict中指出列名字'''#df = pd.DataFrame({'size':train_X,'price':train_y})#df.to_csv('C:/Users/zhangwei/Desktop/Machinelearning/house_price1.csv')'''    由于train_X是一个一唯的list(会被视为一个样本)，需要转化为n_samples形式的二维形式    如果是在csv文件中，读取过程直接转化为此形式    也可以用np.array(train_X)把train转化为 此形式(此方法返回副本，重新赋值给train_X即可)'''#原始的X,y# train_y = [6450.0,7450.0,8450.0,9450.0,11450.0,15450.0,18450.0]# train_X = [150.00, 200.0, 250.0, 300.0, 350.0, 400.0, 600.0]# train_X = np.array(train_X).reshape(-1,1)# print(train_X)#从csv获取数据train_X,train_y = get_data('C:/Users/zhangwei/Desktop/Machinelearning/house_price1.csv')# print(train_X)# print(train_y)'''    intercept:截距    coef:系数    predict_value：预测结果'''def get_predict(train_X,train_y,test_X):    regr = linear_model.LinearRegression()    regr.fit(train_X,train_y)    predict_result = regr.predict(test_X)    predict = {}    predict['intercept'] = regr.intercept_    predict['coef'] = regr.coef_    predict['predict_value'] = predict_result    return predict#对size == 700的房屋进行预测test_X = 700predict = get_predict(train_X,train_y,test_X)print("Intercept value :",predict['intercept'])print("coefficent :",predict['coef'])print('Predicted value:',predict['predict_value'])'''    Intercept value : 1771.80851064    coefficent : [ 28.77659574]    Predicted value: [ 21915.42553191]''''''    绘制预测结果图'''def show_linear_result(train_X,train_y):    regr = linear_model.LinearRegression()    regr.fit(train_X,train_y)    plt.scatter(train_X,train_y,color = "blue")    plt.plot(train_X,regr.predict(train_X),color='red',linewidth=4)    plt.xlabel('house size')    plt.ylabel('house price')    #plt.xticks(())    #plt.yticks(())    plt.show()show_linear_result(train_X,train_y)

预测结果：

参考：

两部电影预测观众喜好：

import csvimport sysimport matplotlib.pyplot as pltimport numpy as npimport pandas as pdfrom sklearn import datasets, linear_model# train_flash_X = [1,2,3,4,5,6,7,8,9]# train_flahs_y = [4.83,4.27,3.59,3.53,3.46,3.73,3.47,4.34,4.66]# train_arrow_X = [1,2,3,4,5,6,7,8,9]# train_arrow_y = [2.84,2.32,2.55,2.49,2.73,2.6,2.64,3.92,3.06]# df = pd.DataFrame({'flash_episod':train_flash_X,'flash_us_viewers':train_flahs_y,'arrow_episod':train_arrow_X,'arrow_us_viewers':train_arrow_y})# df.to_csv('C:/Users/zhangwei/Desktop/Machinelearning/TVshow.csv')# print(train_flash_X)# print(train_flahs_y)# print(train_arrow_X)# print(train_arrow_y)#def get_data(file_name):    data = pd.read_csv(file_name)    train_flash_X = []    train_flash_y = []    train_arrow_X = []    train_arrow_y = []    for x1,x2,x3,x4 in zip(data['flash_episod'],data['flash_us_viewers'],data['arrow_episod'],data['arrow_us_viewers']):        train_flash_X.append([float(x1)])        train_flash_y.append([float(x2)])        train_arrow_X.append([float(x3)])        train_arrow_y.append([float(x4)])    return train_flash_X,train_flash_y,train_arrow_X,train_arrow_y#data = get_data('C:/Users/zhangwei/Desktop/Machinelearning/TVshow.csv')x1,y1,x2,y2 = get_data('C:/Users/zhangwei/Desktop/Machinelearning/TVshow.csv')def more_viewers(x1,y1,x2,y2):    regr1 = linear_model.LinearRegression()    regr1.fit(x1,y1)    predict_value1 = regr1.predict(10)    regr2 = linear_model.LinearRegression()    regr2.fit(x2,y2)    predict_value2 = regr2.predict(10)    print(predict_value1)    print(predict_value2)    if predict_value1 > predict_value2:        print('The Flash tv show will have more viewers for the next week~')    else:        print('The Tv show arrow will have more viewers for the next week~')more_viewers(x1,y1,x2,y2)

数据：