[GdT19] Community-based Recommendations on Twitter: Avoiding The Filter Bubble

Conférence Internationale avec comité de lecture : International Conference on Web Information Systems Engineering (WISE), October 2019, pp.1-15, Hong-Kong,
Résumé: Due to their success, social network platforms are consid-ered today as a major communication mean. In order to increase userengagement, they rely on recommender systems to personalize individualexperience by filtering messages according to user interest and/or neigh-borhood. However some recent results exhibit that this personalizationof content might increase theecho chambereffect and createfilterbub-bles. These filter bubbles restrain the diversity of opinions regarding therecommended content. In this paper, we first realize a thorough study ofcommunities on a large Twitter dataset to quantify how recommendersystems affect users’ behavior and create filter bubbles. Then we proposetheCommunity Aware Model(CAM) to counter the impact of differentrecommender systems on information consumption. Our results showthatfilterbubblesconcern up to 10% of users and our model based onsimilarities between communities enhance recommender systems.


@inproceedings {
title="{Community-based Recommendations on Twitter: Avoiding The Filter Bubble}",
author=" Q. Grossetti and C. du Mouza and N. Travers ",
booktitle="{International Conference on Web Information Systems Engineering (WISE)}",
address=" Hong-Kong",