数据集：银行客户信息

数据集：银行客户信息

# 导包
import pandas as pd
import numpy as np

# 导入数据
train_data = pd.read_csv('modified_bank-additional-train.csv')
test_data = pd.read_csv('modified_bank-additional-test.csv')

# 显示数据维数
train_data.shape

(7873, 21)

test_data.shape

(3964, 21)

# 查看数据集的前4个特征、后2个特征和标签（SalePrice）
train_data.iloc[0:4, [0, 1, 2, 3, -3, -2, -1]]

	age	job	marital	education	euribor3m	nr.employed	y
0	56	services	married	high.school	4.857	5191.0	no
1	54	retired	married	basic.9y	4.857	5191.0	no
2	35	blue-collar	married	basic.6y	4.857	5191.0	no
3	39	management	single	basic.9y	4.857	5191.0	no

# 除去最后一列的特征数据集
all_features = pd.concat((train_data.iloc[:, :-1], test_data.iloc[:, :-1]))
all_features.shape

(11837, 20)

# 查看数据集基本信息
all_features.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 11837 entries, 0 to 3963
Data columns (total 20 columns):
 #   Column          Non-Null Count  Dtype  
---  ------          --------------  -----  
 0   age             11837 non-null  int64  
 1   job             11837 non-null  object 
 2   marital         11837 non-null  object 
 3   education       11837 non-null  object 
 4   default         11837 non-null  object 
 5   housing         11837 non-null  object 
 6   loan            11837 non-null  object 
 7   contact         11837 non-null  object 
 8   month           11837 non-null  object 
 9   day_of_week     11837 non-null  object 
 10  duration        11837 non-null  int64  
 11  campaign        11837 non-null  int64  
 12  pdays           11837 non-null  int64  
 13  previous        11837 non-null  int64  
 14  poutcome        11837 non-null  object 
 15  emp.var.rate    11837 non-null  float64
 16  cons.price.idx  11837 non-null  float64
 17  cons.conf.idx   11837 non-null  float64
 18  euribor3m       11837 non-null  float64
 19  nr.employed     11837 non-null  float64
dtypes: float64(5), int64(5), object(10)
memory usage: 1.9+ MB

# 查看空值信息
all_features.isnull().sum()

age               0
job               0
marital           0
education         0
default           0
housing           0
loan              0
contact           0
month             0
day_of_week       0
duration          0
campaign          0
pdays             0
previous          0
poutcome          0
emp.var.rate      0
cons.price.idx    0
cons.conf.idx     0
euribor3m         0
nr.employed       0
dtype: int64

# 查看重复值信息
all_features.duplicated().sum()

1

# 预处理数据
numeric_features = all_features.dtypes[all_features.dtypes != 'object'].index

# 非object类型数据标准化处理
all_features[numeric_features] = all_features[numeric_features].apply(lambda x: (x - x.mean()) / (x.std()))

# 标准化后，每个特征的均值变为0，所以可以直接用0来填充
all_features[numeric_features] = all_features[numeric_features].fillna(0)

# 离散数值转成指示特征（dummy_na=True将缺失值也当作合法的特征值并为其创建指示特征）
all_features = pd.get_dummies(all_features, dummy_na=True)
all_features.shape

(11837, 73)

# 转换为numpy
n_train = train_data.shape[0]

# 拆分数据集
train_features = np.array(all_features[:n_train].values)
test_features = np.array(all_features[n_train:].values)
train_labels = np.array(train_data.y.values)
test_labels = np.array(test_data.y.values)

# 支持向量机预测
from sklearn.svm import SVC
# from sklearn.linear_model import SGDClassifier
# from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import accuracy_score
clf = SVC(kernel='linear', C=1)
clf.fit(train_features, train_labels)
predictions = clf.predict(test_features)

from sklearn.metrics import accuracy_score
from sklearn.metrics import classification_report
print(classification_report(test_labels, predictions))
print("AC", accuracy_score(test_labels, predictions))

              precision    recall  f1-score   support

          no       0.90      0.86      0.88      2420
         yes       0.80      0.85      0.82      1544

    accuracy                           0.86      3964
   macro avg       0.85      0.86      0.85      3964
weighted avg       0.86      0.86      0.86      3964

AC 0.8572149344096872