FRAUD PATTERN DETECTION BY MOTIF DISCOVERY IN RETAIL BUSINESS

  • Phạm Ngọc Quang Anh
  • Vũ Thành Nam
  • Hoàng Văn Đông
  • Lê Anh Ngọc
  • Nguyễn Thị Ngọc Anh
Keywords: Fraud detection, motif discovery, pattern recognition, machine learning

Abstract

Bad customers do fraud behavior in financial transactions is cause of economic losses and the threat for companies and organizations. Transactions exploded recently because of the development of mobile, online transactions in whole the world. Therefore, transaction processing and detecting anomalous behavior from hundreds of thousands of transactions with various types of behavior are no longer suitable for manual processing. In this paper, motif discovery for time series and anomaly detection by machine learning are proposed. A model that identifies fraudulent behavior patterns and classifies objects in the account level anomaly detection problem is modeled. The proposed model is experimented and then used to discover anomalies in retail activity data. Experimentally, the model has an accuracy of F1 of 75%

điểm /   đánh giá
Published
2022-11-25
Section
Bài viết