Abstract
User interactions with recommender systems (RSs) are affected by user selection bias, e.g., users are more likely to rate popular items (popularity bias) or items that they expect to enjoy beforehand (positivity bias). Methods exist for mitigating the effects of selection bias in user ratings on the evaluation and optimization of RSs. However, these methods treat selection bias as static, despite the fact that the popularity of an item may change drastically over time and the fact that user preferences may also change over time. We focus on the age of an item and its effect on selection bias and user preferences. Our experimental analysis reveals that the rating behavior of users on the MovieLens dataset is better captured by methods that consider effects from the age of item on bias and preferences. We theoretically show that in a dynamic scenario in which both the selection bias and user preferences are dynamic, existing debiasing methods are no longer unbiased. To address this limitation, we introduce DebiAsing in the dyNamiC scEnaRio (DANCER), a novel debiasing method that extends the inverse propensity scoring debiasing method to account for dynamic selection bias and user preferences. Our experimental results indicate that DANCER improves rating prediction performance compared to debiasing methods that incorrectly assume that selection bias is static in a dynamic scenario. To the best of our knowledge, DANCER is the first debiasing method that accounts for dynamic selection bias and user preferences in RSs.
Highlights
User interactions with recommender systems (RSs) are subject to selection bias, as a consequence of the selective behavior of users and of the fact that RSs actively restrict the items from which a user can choose [32, 34, 37, 41, 43]
matrix factorization (MF) by modelling the effect of item-age via element-wise multiplication: pu,i,t =σ (pTu). (7) time-aware tensor factorization (TTF)++: We propose a variation on TTF that models the effect via summation instead: pu,i,t =σ (pTu). (8) Time-aware matrix & tensor factorization (TMTF): Lastly, we propose a novel integration of time-aware matrix factorization (TMF) with TTF++: pu,i,t =
Under both splitting strategies, the time-aware methods TMF, TTF++ and TMTF are significantly more accurate than Pop and MF, which assume that selection bias is static, while MF outperforms Constant, which assumes no bias
Summary
User interactions with recommender systems (RSs) are subject to selection bias, as a consequence of the selective behavior of users and of the fact that RSs actively restrict the items from which a user can choose [32, 34, 37, 41, 43]. A typical form of selection bias in RSs is popularity bias: popular items are often overrepresented in interaction logs because users are more likely to rate them [7, 37, 43]. To correct for selection bias in interaction data from RSs, the task of debiased recommendation has been proposed. A widelyadopted method for this task makes use of inverse propensity scoring (IPS), a causal inference technique [24], and integrates it in the learning process of rating-prediction for recommendation [10, 22, 27, 41]. It estimates the probability of a rating to be observed in the dataset, and inversely weights ratings according to these probabilities so that in expectation each user-item pair is represented
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
