Doubly High-Dimensional Contextual Bandits: An Interpretable Model for Joint Assortment-Pricing

TL;DR

This paper introduces a joint assortment-pricing model using high-dimensional contextual bandits with a low-rank structure, achieving interpretable insights and significant revenue gains in real-world retail scenarios.

Contribution

It proposes a novel doubly high-dimensional bandit model with a low-rank interaction structure, combining interpretability with computational efficiency and theoretical regret bounds.

Findings

Lower regret than state-of-the-art methods in simulations

At least three-fold revenue or profit gains in case studies

Model captures interactions via interpretable latent factors

Abstract

Key challenges in running a retail business include how to select products to present to consumers (the assortment problem), and how to price products (the pricing problem) to maximize revenue or profit. Instead of considering these problems in isolation, we propose a joint approach to assortment-pricing based on contextual bandits. Our model is doubly high-dimensional, in that both context vectors and actions are allowed to take values in high-dimensional spaces. In order to circumvent the curse of dimensionality, we propose a simple yet flexible model that captures the interactions between covariates and actions via a (near) low-rank representation matrix. The resulting class of models is reasonably expressive while remaining interpretable through latent factors, and includes various structured linear bandit and pricing models as particular cases. We propose a computationally tractable procedure that combines an exploration/exploitation protocol with an efficient low-rank matrix estimator, and we prove bounds on its regret. Simulation results show that this method has lower regret than state-of-the-art methods applied to various standard bandit and pricing models. Real-world case studies on the assortment-pricing problem, from an industry-leading instant noodles company to an emerging beauty start-up, underscore the gains achievable using our method. In each case, we show at least three-fold gains in revenue or profit by our bandit method, as well as the interpretability of the latent factor models that are learned.