A Predict-Then-Optimize Customer Allocation Framework for Online Fund Recommendation

2025-03-05

AI Summary

Online fund recommendation faces a challenging client-fund matching problem under multiple hard constraints (e.g., regulatory compliance, resource quotas), where conventional recommender systems struggle to simultaneously maximize returns and satisfy all constraints. This paper proposes PTOFA, a two-stage Prediction–Optimization framework: (1) a deep behavioral modeling stage predicts users’ expected returns; (2) an optimization stage allocates fund exposures via integer linear programming or heuristic solvers, strictly enforcing all hard constraints. To our knowledge, this is the first work to systematically introduce the end-to-end prediction–optimization paradigm into fund matching—yielding interpretable, controllable, and constraint-feasible allocations. Evaluated on real-world platform data, PTOFA achieves a 12.7% ROI improvement; A/B testing shows a 9.3% daily increase in transaction volume, with 100% constraint satisfaction. The system has been deployed at a leading wealth management platform.

Technology Category

• Search and Optimization: Other Foundations of Search & Optimization
• Constraint Satisfaction and Optimization: Constraint Optimization
• Machine Learning: Optimization

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Economics and fairness of platforms and recommendation systems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

With the rapid growth of online investment platforms, funds can be distributed to individual customers online. The central issue is to match funds with potential customers under constraints. Most mainstream platforms adopt the recommendation formulation to tackle the problem. However, the traditional recommendation regime has its inherent drawbacks when applying the fund-matching problem with multiple constraints. In this paper, we model the fund matching under the allocation formulation. We design PTOFA, a Predict-Then-Optimize Fund Allocation framework. This data-driven framework consists of two stages, i.e., prediction and optimization, which aim to predict expected revenue based on customer behavior and optimize the impression allocation to achieve the maximum revenue under the necessary constraints, respectively. Extensive experiments on real-world datasets from an industrial online investment platform validate the effectiveness and efficiency of our solution. Additionally, the online A/B tests demonstrate PTOFA's effectiveness in the real-world fund recommendation scenario.

Problem

• Match funds with customers under constraints.
• Overcome drawbacks of traditional recommendation systems.
• Maximize revenue through predictive and optimization stages.

Innovation

• Predict-Then-Optimize Fund Allocation framework
• Two-stage data-driven approach
• Revenue prediction and impression optimization