Community-Based Recommendations on Twitter: Avoiding the Filter Bubble

Abstract

Due to their success, social network platforms are considered today as a major communication mean. In order to increase user engagement, they rely on recommender systems to personalize individual experience by filtering messages according to user interest and/or neighborhood. However some recent results exhibit that this personalization of content might increase the echo chamber effect and create filter bubbles. These filter bubbles restrain the diversity of opinions regarding the recommended content. In this paper, we first realize a thorough study of communities on a large Twitter dataset to quantify how recommender systems affect users' behavior and create filter bubbles. Then we propose the Community Aware Model (CAM) to counter the impact of different recommender systems on information consumption. Our results show that filter bubbles concern up to 10% of users and our model based on similarities between communities enhance recommender systems.

Network polarization, filter bubbles, and echo chambers: An annotated review of measures, models, and case studies

“…A number of studies stand that social networks are in themselves causally sufficient to promote echo chambers due to their structure [94], size [85], lack of trust [79], or the embedded recommendation systems [62]. Therefore, they argue that promoting features such as transparency [89], trust [79], or improved recommendation systems [62] can reduce polarization.…”

“…In particular, some studies indicate that content personalization produced by recommendation systems may increase the echo chamber effect and create filter bubbles [62]. Therefore, some authors claim that modifying recommendation systems may reduce network polarization [4,62,93].…”

“…Grossetti et al [62,63] studied communities on a large Twitter dataset to quantify how recommendation systems affect users' behavior, and how content personalization can increase the echo chamber effect and create filter bubbles. A preliminary study was conducted to detect a filter bubble effect on users' community profiles, proposing a community-aware model whose objective is to reduce the filter-bubble impact.…”

Mitigating Filter Bubbles within Deep Recommender Systems

“…Recently, Grosetti et. al [5] quantified how standard recommender systems can affect users' behaviors and amplify filter bubbles with Twitter data. They create profiles for users based on interaction histories, and then generate recommendations for those users using a recommender system.…”

“…A mitigation strategy is proposed in [5] by re-ranking the outputs from the recommender system algorithms and minimizing an additional objective function-the distance between the user's profile vector and a community score vector generated from the recommendations. The authors were able to mitigate the filter bubble in the Graphjet, CF, and SimGraph recommender algorithms (as defined by Gini coefficients).…”