RecTwitter: A Semantic Recommendation System for Twitter Users
Twitter;RECTWITTER
Twitter is one of the most popular microblogging services today, which allows users to share images, links, texts, etc., as well as follow or not to follow preferred accounts, which are also managed by other users. Public opinion has to do with the differences in volume and volume of publication. In view of this, it is necessary to adopt intelligent schemas to identify and filter accounts that publish content similar to the interests of the target user. This dissertation offers a comment system for a subject with rules of conversation, which leads us to analyze the accounts of a user, and recommends those that are discontinued and new ones that are followed in a row. As rules work as recommendation engines, people are shaped through an interaction between Twitter users as they are modeled through a domain ontology. To evaluate the proposed model, experiments were carried out with real users and comparisons with works related to the state of the art. The results of the online experiments indicate that 76% of users were approved as a reference. In relation to the results obtained in the off-line experiments, one of the best models of the state of the art.