Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Financial Risk Assessment via Long-term Payment Behavior Sequence Folding

Yiran Qiao, Yateng Tang, Xiang Ao, Qi Yuan, Ziming Liu, Chen Shen, Xuehao Zheng

TL;DR

The results demonstrate that folding long behavior sequences based on internal behavioral cues effectively models long-term patterns and changes, thereby generating more accurate user financial profiles for practical applications.

Abstract

Online inclusive financial services encounter significant financial risks due to their expansive user base and low default costs. By real-world practice, we reveal that utilizing longer-term user payment behaviors can enhance models' ability to forecast financial risks. However, learning long behavior sequences is non-trivial for deep sequential models. Additionally, the diverse fields of payment behaviors carry rich information, requiring thorough exploitation. These factors collectively complicate the task of long-term user behavior modeling. To tackle these challenges, we propose a Long-term Payment Behavior Sequence Folding method, referred to as LBSF. In LBSF, payment behavior sequences are folded based on merchants, using the merchant field as an intrinsic grouping criterion, which enables informative parallelism without reliance on external knowledge. Meanwhile, we maximize the utility of payment details through a multi-field behavior encoding mechanism. Subsequently, behavior aggregation at the merchant level followed by relational learning across merchants facilitates comprehensive user financial representation. We evaluate LBSF on the financial risk assessment task using a large-scale real-world dataset. The results demonstrate that folding long behavior sequences based on internal behavioral cues effectively models long-term patterns and changes, thereby generating more accurate user financial profiles for practical applications.

Financial Risk Assessment via Long-term Payment Behavior Sequence Folding

TL;DR

The results demonstrate that folding long behavior sequences based on internal behavioral cues effectively models long-term patterns and changes, thereby generating more accurate user financial profiles for practical applications.

Abstract

Online inclusive financial services encounter significant financial risks due to their expansive user base and low default costs. By real-world practice, we reveal that utilizing longer-term user payment behaviors can enhance models' ability to forecast financial risks. However, learning long behavior sequences is non-trivial for deep sequential models. Additionally, the diverse fields of payment behaviors carry rich information, requiring thorough exploitation. These factors collectively complicate the task of long-term user behavior modeling. To tackle these challenges, we propose a Long-term Payment Behavior Sequence Folding method, referred to as LBSF. In LBSF, payment behavior sequences are folded based on merchants, using the merchant field as an intrinsic grouping criterion, which enables informative parallelism without reliance on external knowledge. Meanwhile, we maximize the utility of payment details through a multi-field behavior encoding mechanism. Subsequently, behavior aggregation at the merchant level followed by relational learning across merchants facilitates comprehensive user financial representation. We evaluate LBSF on the financial risk assessment task using a large-scale real-world dataset. The results demonstrate that folding long behavior sequences based on internal behavioral cues effectively models long-term patterns and changes, thereby generating more accurate user financial profiles for practical applications.

Paper Structure

This paper contains 33 sections, 6 equations, 5 figures, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Financial Risk Assessment
  4. User Behavior Sequence Modeling for Fraud Detection
  5. Transformer-based Long Sequence Modeling
  6. Problem Formulation
  7. Business Setting
  8. Problem Statement
  9. Methodology
  10. Overview of LBSF
  11. Long-term Payment Behavior Reorganizing
  12. Multi-field Behavior Encoding
  13. Hierarchical Sequence Melding
  14. Melding within a merchant
  15. Melding across merchants
  16. ...and 18 more sections

Figures (5)

  • Figure 1: The performance of Transformer and BERT4Rec on real-world datasets of 45 days, 90 days, and 180 days, measured by AUC.
  • Figure 2: The architecture of LBSF.
  • Figure 3: The illustration of long-term payment behavior reorganizing at merchant level on a short example sequence.
  • Figure 4: Case study of defaulter Alice, whose four top-ranked merchants are a consumer finance app, an e-commerce app, a high-end restaurant, and public transportation.
  • Figure 5: Case study of defaulter Bob, whose three top-ranked merchants are a live-streaming app, a food delivery app, and a game app.