Dynamic collaborative filtering Thompson Sampling for cross-domain advertisements recommendation

TL;DR

This paper introduces DCTS, a novel model that enhances cross-domain ad recommendation by transferring knowledge among models using similarities and temporal user dynamics, leading to improved click-through rates.

Contribution

The paper proposes DCTS, a simple yet effective model that leverages similarities and temporal dynamics for knowledge transfer in multi-domain recommender systems.

Findings

DCTS improves CTR by 9.7% over state-of-the-art models.

Knowledge transfer and temporal dynamics enhance model performance.

Optimal hyper-parameters significantly impact CTR maximization.

Abstract

Recently online advertisers utilize Recommender systems (RSs) for display advertising to improve users' engagement. The contextual bandit model is a widely used RS to exploit and explore users' engagement and maximize the long-term rewards such as clicks or conversions. However, the current models aim to optimize a set of ads only in a specific domain and do not share information with other models in multiple domains. In this paper, we propose dynamic collaborative filtering Thompson Sampling (DCTS), the novel yet simple model to transfer knowledge among multiple bandit models. DCTS exploits similarities between users and between ads to estimate a prior distribution of Thompson sampling. Such similarities are obtained based on contextual features of users and ads. Similarities enable models in a domain that didn't have much data to converge more quickly by transferring knowledge. Moreover, DCTS incorporates temporal dynamics of users to track the user's recent change of preference. We first show transferring knowledge and incorporating temporal dynamics improve the performance of the baseline models on a synthetic dataset. Then we conduct an empirical analysis on a real-world dataset and the result showed that DCTS improves click-through rate by 9.7% than the state-of-the-art models. We also analyze hyper-parameters that adjust temporal dynamics and similarities and show the best parameter which maximizes CTR.