Research

Working in Progress


Abstract: I explore how a platform provides product recommendations with discriminatory accuracy based on buyers’ preference intensities. The platform balances accurate recommendations that foster trade with biased ones favoring high-margin products. When buyers’ types are observable, the platform’s optimal policy is extremely contrary to buyers’ preferences, i.e., it offers accurate recommendations only to ”picky” buyers who benefit the least, while steering less ”picky” buyers toward high-margin products despite their greater eagerness for suitable matches. In contrast, when the platform must screen buyers, it mimics a system where buyers self-select their preferred accuracy levels, achieving the most efficient allocation of recommendation accuracy.




Conference Presentation

Optimal Allocation of Accuracy in Personalized Product Recommendation, Econometric Society Australasian Meetings. Monash University, Caulfield Campus, Melbourne, Australia, December 2024.




Earlier Publications

Bayesian truncation errors in equations of state of nuclear matter with chiral nucleon-nucleon potentials, joint with Jinniu Hu and Ying Zhang, Physics Letters B 798 (2019): 134982.


How New Rural Cooperative Medical System Affect Labor Supply of Middle-Aged and Elderly Population in Rural Areas: Based on Analysis of Dynamic Stochastic Model, joint with Zhaona, Finance and Economics(in Chinese)