Multi-Instance Learning Based Anomaly Detection Method for Sequence Data with Application to the Credit Card Delinquency Risk Control
Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Keywords: Anomaly Detection, Complex Sequential Data Analysis, Credit Card Delinquency Risk Control, Multi-Instance Learning
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
Abstract: Anomaly detection in sequence data is widely applicable to many fields and has significant commercial value to the financial industry. The focus of this paper is its utility as means to control credit card delinquency risk. Transactions that deviate from the typical data sequence are a common precursor of payment difficulty. Current detection methods do not effectively use transaction data to detect abnormal transactions. This makes it difficult to control the overdue payment risk. We propose a Multi-Instance Learning based anomaly detection (MILAD) method with well designed learning networks to address this problem. MILAD analyze users’ monthly transactions and payment history, and detect exceptions through well designed deep learning networks. By comparing the performance of MILAD and DAGMM, which is currently the most commonly used unsupervised deep learning algorithm for credit card risk control, MILAD best controls overdue risk by utilizing both transaction and payment information.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 182
Loading