Loading...
Paper Type
Complete
Abstract
Twitter is a microblog that contains large amounts of users who contribute with messages for a wide variety of real-world events. It is possible to identify users who share the same interests using the messages published in their timeline. However, those users can stop publishing interesting content anytime, thus finding any interesting content manually is a hard task. Taken into account users' access on Twitter, he may lose important tweets, or only a part of all tweets will be relevant to him. Indeed, we can observe that it is important to develop automated mechanisms to filter out these messages. In this project, we propose a semantic recommendation system based on SWRL rules to recommend accounts to be followed or unfollowed. To evaluate the recommendations, we conducted an experiment with real users. The results show that 80% of the recommendations were generated to unfollow and 20% to follow some account.
Recommended Citation
de Souza, Paulo Roberto and Durao, Frederico Araujo, "RecTwitter: A Rule-Based Semantic Recommender System for Twitter Users" (2020). AMCIS 2020 Proceedings. 25.
https://aisel.aisnet.org/amcis2020/ai_semantic_for_intelligent_info_systems/ai_semantic_for_intelligent_info_systems/25
RecTwitter: A Rule-Based Semantic Recommender System for Twitter Users
Twitter is a microblog that contains large amounts of users who contribute with messages for a wide variety of real-world events. It is possible to identify users who share the same interests using the messages published in their timeline. However, those users can stop publishing interesting content anytime, thus finding any interesting content manually is a hard task. Taken into account users' access on Twitter, he may lose important tweets, or only a part of all tweets will be relevant to him. Indeed, we can observe that it is important to develop automated mechanisms to filter out these messages. In this project, we propose a semantic recommendation system based on SWRL rules to recommend accounts to be followed or unfollowed. To evaluate the recommendations, we conducted an experiment with real users. The results show that 80% of the recommendations were generated to unfollow and 20% to follow some account.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.