face-to-face credit card transactions dataset

本研究使用的数据集名为‘面对面信用卡交易数据集’，由工业合作伙伴提供，涵盖了2015年3月1日至5月31日比利时所有信用卡的匿名交易记录。数据集包含92天的交易数据，旨在通过分析交易行为的变化来量化数据集偏移，以改进信用卡欺诈检测。数据集的创建过程涉及对每日交易进行分类，并构建距离矩阵以量化不同日期间的交易行为差异。该数据集主要应用于信用卡欺诈检测领域，通过识别和适应交易行为的变化，提高欺诈检测系统的准确性和适应性。

The dataset utilized in this study is named the "Face-to-Face Credit Card Transaction Dataset", which was provided by industrial partners and covers anonymous transaction records of all credit cards in Belgium spanning from March 1 to May 31, 2015. This dataset contains 92 days of transaction data, and its purpose is to quantify dataset shift by analyzing changes in transaction behaviors so as to advance credit card fraud detection. The process of creating this dataset involves classifying daily transactions and constructing distance matrices to quantify the differences in transaction behaviors across different days. This dataset is primarily applied in the credit card fraud detection domain, where it can improve the accuracy and adaptability of fraud detection systems by identifying and adapting to shifts in transaction behaviors.