import pandas as pd import numpy as np import seaborn as sns import matplotlib.pyplot as plt from sklearn.ensemble import RandomForestRegressor from sklearn.model_selection import train_test_split from sklearn.linear_model import LinearRegression from sklearn.tree import DecisionTreeRegressor from sklearn.metrics import mean_squared_error from sklearn.neural_network import MLPRegressor from sklearn.svm import SVR

读取数据

df_ads = pd.read_csv('易速鲜花微信软文.csv') print(df_ads.head(10))

数据清洗

df_ads.isna().sum() # NaN出现的次数 print(df_ads.isna().sum()) df_ads = df_ads.dropna() # 删除NaN值 print(df_ads)

数据分析

plt.plot(df_ads['点赞数'], df_ads['浏览量'], 'r.', label='Training data') plt.xlabel('点赞数') plt.ylabel('浏览量') plt.legend() plt.show() data = pd.concat([df_ads['浏览量'], df_ads['热度指数']], axis=1) # 浏览量和热度指数 fig = sns.boxplot(x='热度指数', y='浏览量', data=data) # 用seaborn的箱线图画图 fig.axis(ymin=0, ymax=800000); #设定y轴坐标 plt.show()

特征工程

X = df_ads[['点赞数']] y = df_ads['浏览量']

数据拆分

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)

模型构建

model = RandomForestRegressor() model.fit(X_train, y_train)

预测结果

y_pred = model.predict(X_test)

模型评估

print('随机森林模型评估：') print('均方误差：%.2f' % mean_squared_error(y_test, y_pred)) print('R2得分：%.2f' % model.score(X_test, y_test))

模型构建

model = DecisionTreeRegressor() model.fit(X_train, y_train)

预测结果

y_pred = model.predict(X_test)

模型评估

print('决策树模型评估：') print('均方误差：%.2f' % mean_squared_error(y_test, y_pred)) print('R2得分：%.2f' % model.score(X_test, y_test))

模型构建

model = LinearRegression() model.fit(X_train, y_train)

预测结果

y_pred = model.predict(X_test)

模型评估

print('线性回归模型评估：') print('均方误差：%.2f' % mean_squared_error(y_test, y_pred)) print('R2得分：%.2f' % model.score(X_test, y_test))

model_svr = SVR(kernel='rbf', C=1e3, gamma=0.1) model_svr.fit(X_train, y_train) y_pred_svr = model_svr.predict(X_test) df_ads_pred_svr = X_test.copy() df_ads_pred_svr['浏览量真值'] = y_test df_ads_pred_svr['浏览量预测值'] = y_pred_svr df_ads_pred_svr print('支持向量机预测集评分：', model_svr.score(X_test, y_test)) print('支持向量机训练集评分：', model_svr.score(X_train, y_train))

model_mlp = MLPRegressor(hidden_layer_sizes=(100,50,10), max_iter=1000, alpha=0.001, solver='adam', verbose=0, random_state=21) model_mlp.fit(X_train, y_train) y_pred_mlp = model_mlp.predict(X_test) df_ads_pred_mlp = X_test.copy() df_ads_pred_mlp['浏览量真值'] = y_test df_ads_pred_mlp['浏览量预测值'] = y_pred_mlp df_ads_pred_mlp print('神经网络预测集评分：', model_mlp.score(X_test, y_test)) print('神经网络训练集评分：', model_mlp.score(X_train, y_train))

加入两个新的图表，一个是点赞数和浏览量的饼图，一个是热度指数和浏览量的饼图

饼图1

labels = ['0-1000', '1000-3000', '3000-5000', '5000-10000', '>10000'] df_ads['点赞数分布'] = pd.cut(df_ads['点赞数'], bins=[0, 1000, 3000, 5000, 10000, np.inf], labels=labels) df_pie = df_ads.groupby('点赞数分布').size().reset_index(name='counts') fig1, ax1 = plt.subplots() ax1.pie(df_pie['counts'], labels=df_pie['点赞数分布'], autopct='%1.1f%%') ax1.axis('equal') plt.title('点赞数分布') plt.show()

饼图2

labels = ['0-20', '20-40', '40-60', '60-80', '>80'] df_ads['热度指数分布'] = pd.cut(df_ads['热度指数'], bins=[0, 20, 40, 60, 80, np.inf], labels=labels) df_pie = df_ads.groupby('热度指数分布').size().reset_index(name='counts') fig2, ax2 = plt.subplots() ax2.pie(df_pie['counts'], labels=df_pie['热度指数分布'], autopct='%1.1f%%') ax2.axis('equal') plt.title('热度指数分布') plt.show()