要求：
1、根据相关的信息预测通过电话推销，用户是否会在银行进行存款
2、数据bank.csv，约4520条数据，17个属性值 提示： 17个属性，分别是年龄，工作类型，婚姻状况，受教育背景，信用情况，房贷，个人贷款，联系电话是手机还是固定电话，最后联系月份，最后联系日，通话持续时间，本次活动中联系的次数，最后一次接触距离上一次接触的时间，以前的活动中联系的次数，上一次活动成功与否，是否会进行存款
代码实现：
python
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler, OneHotEncoder
from sklearn.compose import ColumnTransformer
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score, f1_score, confusion_matrix, roc_auc_score
import matplotlib.pyplot as plt
import seaborn as sns
plt.rcParams['font.sans-serif'] = ['SimHei']
plt.rcParams['axes.unicode_minus'] = False
#加载本地的数据
file_path = r'D:\课程\数据挖掘\实验八\实验8-bank.csv'
df = pd.read_csv(file_path, sep=';', quotechar='"')
#进行数据预处理
df = df.replace('unknown', np.nan)
for col in ['default', 'housing', 'loan', 'y']:
df[col] = df[col].map({'yes': 1, 'no': 0, 'unknown': np.nan})
#处理缺失值
mode_values = {
'job': df['job'].mode()[0] if not df['job'].mode().empty else 'unknown',
'education': df['education'].mode()[0] if not df['education'].mode().empty else 'unknown',
'contact': df['contact'].mode()[0] if not df['contact'].mode().empty else 'unknown',
'poutcome': df['poutcome'].mode()[0] if not df['poutcome'].mode().empty else 'unknown'
}
df = df.fillna(mode_values)
#标识特征
df['pdays'] = df['pdays'].replace(-1, 999) #未联系过的特殊标记
df['contacted_before'] = np.where(df['pdays'] != 999, 1, 0) #是否曾经联系过
#检查并处理目标变量中的NaN
print(f"目标变量y中的缺失值数量: {df['y'].isna().sum()}")
df = df.dropna(subset=['y'])
#特征编码
categorical_features = ['job', 'marital', 'education', 'contact', 'month', 'poutcome']
numeric_features = ['age', 'balance', 'day', 'duration', 'campaign', 'pdays', 'previous']
preprocessor = ColumnTransformer(
transformers=[
('num', StandardScaler(), numeric_features),
('cat', OneHotEncoder(handle_unknown='ignore', sparse_output=False), categorical_features)
])
X = df.drop('y', axis=1)
y = df['y']
#进行数据集划分
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.2, random_state=42, stratify=y
)
#模型训练
rf = RandomForestClassifier(
n_estimators=100,
max_depth=8,
min_samples_split=5,
class_weight='balanced',
random_state=42
)
X_train_processed = preprocessor.fit_transform(X_train)
rf.fit(X_train_processed, y_train)
#模型评估
X_test_processed = preprocessor.transform(X_test)
y_pred = rf.predict(X_test_processed)
y_proba = rf.predict_proba(X_test_processed)[:, 1]
print(f"准确率: {accuracy_score(y_test, y_pred):.4f}")
print(f"F1分数: {f1_score(y_test, y_pred):.4f}")
print(f"AUC: {roc_auc_score(y_test, y_proba):.4f}")
#混淆矩阵可视化
cm = confusion_matrix(y_test, y_pred)
plt.figure(figsize=(8, 6))
sns.heatmap(cm, annot=True, fmt='d', cmap='Blues',
xticklabels=['不会存款', '会存款'],
yticklabels=['不会存款', '会存款'])
plt.xlabel('预测标签')
plt.ylabel('真实标签')
plt.title('混淆矩阵')
plt.show()
#获取特征名称
cat_encoder = preprocessor.named_transformers_['cat']
cat_features = cat_encoder.get_feature_names_out(categorical_features)
all_features = np.concatenate([numeric_features, cat_features])
#绘制特征重要性
importances = rf.feature_importances_
indices = np.argsort(importances)[::-1][:15]
plt.figure(figsize=(12, 8))
plt.title("Top 15 特征重要性")
plt.barh(range(15), importances[indices][::-1], align='center')
plt.yticks(range(15), all_features[indices][::-1])
plt.xlabel('重要性')
plt.tight_layout()
plt.show()
#精准营销建议
def marketing_recommendation(model, preprocessor, customer_data):
processed_data = preprocessor.transform(customer_data)
proba = model.predict_proba(processed_data)[0][1]
return proba > 0.4
#客户数据
sample_customer = pd.DataFrame({
'age': [42],
'job': ['management'],
'marital': ['married'],
'education': ['tertiary'],
'default': [0],
'balance': [1500],
'housing': [1],
'loan': [0],
'contact': ['cellular'],
'day': [15],
'month': ['may'],
'duration': [300],
'campaign': [2],
'pdays': [100],
'previous': [3],
'poutcome': ['success']
})
if marketing_recommendation(rf, preprocessor, sample_customer):
print("推荐对该客户进行存款营销")
else:
print("不建议对该客户进行存款营销")运行结果：
混淆矩阵：
Top15特征重要性
控制台输出，推荐对该客户进行存款营销
