第八章 分类 案例1：社交网络平台汽车广告精准营销

案例背景

本案例中使用的数据集包含年龄、工资和购买特定汽车的决定。案例中将简单使用KNN算法来预测某个特定的人是否会买车

数据读取

import numpy as np
import os
import matplotlib.pyplot as plt 
import pandas as pd 
import seaborn as sns
from sklearn.model_selection import train_test_split   
from sklearn.preprocessing import StandardScaler 
from matplotlib.colors import ListedColormap
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import confusion_matrix,classification_report,accuracy_score
import warnings
warnings.filterwarnings('ignore') 
plt.style.use('fivethirtyeight')

dataset=pd.read_csv('./Social_Network_Ads.csv')
dataset.head(2)

dataset.describe().T

平均年龄大概37岁，平均工资69742$

f,ax=plt.subplots(1,2,figsize=(18,8))
dataset['Gender'].value_counts().plot.pie(explode=[0,0.05],autopct='%1.1f%%',ax=ax[0],shadow=True)
ax[0].set_title('Purchase by Gender')
ax[0].set_ylabel('Count')
sns.countplot('Gender',data=dataset,ax=ax[1],order=dataset['Gender'].value_counts().index)
ax[1].set_title('Purchase by Gender')
plt.show()

看出男女比较平等（男女数量相当）

分析薪资分布

fig=plt.gcf()
fig.set_size_inches(10,7)
fig=sns.distplot(dataset['EstimatedSalary'],kde=True,bins=50)

我们看到人群的平均工资大概是70000$

f,ax=plt.subplots(1,2,figsize=(18,8))
dataset['Purchased'].value_counts().plot.pie(explode=[0,0.05],autopct='%1.1f%%',ax=ax[0],shadow=True)
ax[0].set_title('Purchase Distribution')
ax[0].set_ylabel('Count')
sns.countplot('Purchased',data=dataset,ax=ax[1],order=dataset['Purchased'].value_counts().index)
ax[1].set_title('Purchase Distribution')
plt.show()

所以数据集中的大多数人都没有买车。
我们的一切努力都应该是为了销售更多的汽车。

模型构建与训练

X=dataset.iloc[:,[2,3]].values
y=dataset.iloc[:,4].values

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

X_train=StandardScaler().fit_transform(X_train)
X_test=StandardScaler().fit_transform(X_test)

classifier=KNeighborsClassifier(n_neighbors=1,metric='minkowski',p=2)
classifier.fit(X_train,y_train)

KNeighborsClassifier(n_neighbors=1)

KNeighborsClassifier(n_neighbors=1)

y_pred=classifier.predict(X_test)

cm=confusion_matrix(y_test,y_pred)
f, ax = plt.subplots(figsize =(5,5))
sns.heatmap(cm,annot = True,linewidths=0.5,linecolor="red",fmt = ".0f",ax=ax)
plt.title("Test for Test Dataset")
plt.xlabel("predicted y values")
plt.ylabel("real y values")
plt.show()

print(accuracy_score(y_test,y_pred))

0.925

模型参数调优

error_rate=[]

for i in range(1,40):
    knn=KNeighborsClassifier(n_neighbors=i)
    knn.fit(X_train,y_train)
    pred_i=knn.predict(X_test)
    error_rate.append(np.mean(pred_i != y_test))

plt.figure(figsize=(10,6))
plt.plot(range(1,40),error_rate,color='blue',linestyle='dashed',marker='o',markerfacecolor='red',markersize=10)
plt.title('Error Rate Vs K Value')
plt.xlabel('K')
plt.ylabel('Error Rate')
plt.show()

classifier=KNeighborsClassifier(n_neighbors=5,metric='minkowski',p=2)
classifier.fit(X_train,y_train)

KNeighborsClassifier()

KNeighborsClassifier()

y_pred=classifier.predict(X_test)

cm=confusion_matrix(y_test,y_pred)
f, ax = plt.subplots(figsize =(5,5))
sns.heatmap(cm,annot = True,linewidths=0.5,linecolor="red",fmt = ".0f",ax=ax)
plt.title("Test for Test Dataset")
plt.xlabel("predicted y values")
plt.ylabel("real y values")
plt.show()

print(accuracy_score(y_test,y_pred))

0.9375