Author(s): Shenggang Li

Originally published on Towards AI.

Clustered contextual bandits and tabular Q-learning with off-policy evaluation on real-world retail logs

Traditional market basket analysis (MBA) explains what tends to co-occur, but it does not decide what to do next. This paper advances from static rules to policy learning — we model shopping as a sequential decision problem and learn which product to recommend next to maximize business value (e.g., margin or revenue).

The article discusses an advancement in market basket analysis through the application of reinforcement learning techniques, particularly segmented contextual bandits and tabular Q-learning, aimed at optimizing product recommendations in a retail context. It describes the construction of a pipeline that learns effective recommendation policies from historical retail transactions, evaluates several methodologies, and demonstrates that the proposed reinforcement learning approaches outperform traditional methods by better adapting to customer contexts and optimizing long-term business value.

Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI