1.数据读取
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
%matplotlib inline
data = pd.read_csv("creditcard.csv")
data.head()
展示数据基本信息,如缺失值,字段类型等等
data.info()
2.数据预处理
#统计不同标签对应的样本数
count_classes = pd.value_counts(data['Class'], sort = True).sort_index()
#通过条形图显示出来
count_classes.plot(kind = 'bar')
plt.title("Fraud class histogram")
plt.xlabel("Class")
plt.ylabel("Frequency")
通过条形图显示,正负样本(class分别为1和0的样本)比例失调,通常解决这一问题的方法有:
- 调整正负样本的权重
- 上采样
- 下采样
#对样本的Amount列进行归一化处理
from sklearn import preprocessing
data['normAmount'] = preprocessing.scale(data['Amount'])
#删除无关列
data = data.drop(['Time', 'Amount'], axis=1)
data.head()
生成下采样样本数据集
#筛选出训练数据与相应的分类标签
X = data.loc[:, data.columns != 'Class']
y = data.loc[:, data.columns == 'Class']
# 筛选出信用卡欺诈样本,并统计样本数量
number_records_fraud = len(data[data.Class == 1])
fraud_indices = np.array(data[data.Class == 1].index)
# 筛选出正常样本,并统计样本数量
normal_indices = data[data.Class == 0].index
# 由于信用卡欺诈的样本数量远远小于正常样本数量,我们这里采取下采样,即删去正常样本,使其样本数量和信用卡欺诈样本数量保持一致
random_normal_indices = np.random.choice(normal_indices, number_records_fraud, replace = False)
random_normal_indices = np.array(random_normal_indices)
# 生成下采样样本
under_sample_indices = np.concatenate([fraud_indices,random_normal_indices])
under_sample_data = data.iloc[under_sample_indices,:]
# 分离下采样样本的数据集及标签集
X_undersample = under_sample_data.loc[:, under_sample_data.columns != 'Class']
y_undersample = under_sample_data.loc[:, under_sample_data.columns == 'Class']
# 打印出正负样本比例
print("Percentage of normal transactions: ", len(under_sample_data[under_sample_data.Class == 0])/len(