Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

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

Xing Tang, Yunpeng Weng, Fuyuan Lyu, Dugang Liu, Xiuqiang He

TL;DR

The paper tackles the problem of matching funds to customers on online investment platforms under exposure and risk constraints. It introduces the two-stage Predict-Then-Optimize Fund Allocation (PTOFA) framework, combining a counterfactual, full-space revenue predictor with a fast heuristic allocator to maximize total revenue. Key contributions include recasting fund matching as allocation, developing a full-space revenue prediction loss, and delivering a scalable optimization algorithm that approaches integer-programming performance while remaining tractable at large scale. Empirical results from offline datasets and online A/B tests show notable improvements in conversion, revenue per exposure, and overall allocation efficiency, validating PTOFA's practical value for real-world platforms like LiCaiTong.

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.

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

TL;DR

The paper tackles the problem of matching funds to customers on online investment platforms under exposure and risk constraints. It introduces the two-stage Predict-Then-Optimize Fund Allocation (PTOFA) framework, combining a counterfactual, full-space revenue predictor with a fast heuristic allocator to maximize total revenue. Key contributions include recasting fund matching as allocation, developing a full-space revenue prediction loss, and delivering a scalable optimization algorithm that approaches integer-programming performance while remaining tractable at large scale. Empirical results from offline datasets and online A/B tests show notable improvements in conversion, revenue per exposure, and overall allocation efficiency, validating PTOFA's practical value for real-world platforms like LiCaiTong.

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.

Paper Structure

This paper contains 10 sections, 8 equations, 4 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Formulation of Fund Allocation
  3. Methodology
  4. Experiments
  5. Offline Experiments
  6. Result from online A/B testings
  7. Related Work
  8. Generalized Recommendation and Fund Recommendation
  9. Predict-then-Optimize Problem
  10. Conclusion

Figures (4)

  • Figure 1: An illustration of fund matching process on an online investment platform.
  • Figure 2: Distribution of logarithmic transaction value.
  • Figure 3: The existing solution and our idea.
  • Figure 4: Comparison of different allocation strategies.