Exploring Popularity Bias in Session-based Recommendation
This work addresses bias in recommender systems for users in session-based contexts, but it is incremental as it adapts existing methods to a new setup.
The study tackled the problem of popularity bias in session-based recommendation systems by extending propensity analysis to this setup and adapting its calculation to session-specific characteristics, achieving promising results compared to original models.
Existing work has revealed that large-scale offline evaluation of recommender systems for user-item interactions is prone to bias caused by the deployed system itself, as a form of closed loop feedback. Many adopt the \textit{propensity} concept to analyze or mitigate this empirical issue. In this work, we extend the analysis to session-based setup and adapted propensity calculation to the unique characteristics of session-based recommendation tasks. Our experiments incorporate neural models and KNN-based models, and cover both the music and the e-commerce domain. We study the distributions of propensity and different stratification techniques on different datasets and find that propensity-related traits are actually dataset-specific. We then leverage the effect of stratification and achieve promising results compared to the original models.