Cluster Aware Personalized News Recommendation System Using a Global News Graph

With the increasing use of online news services and the production of a large amount of news daily, finding news content that matches user interests has become an important challenge. Personalized news recommendation systems, by mod-eling users' behavior and interests, can help us to respond to this challenge. In this study, a new method for personalized news recommendations based on clustering and graph enhancement has been introduced, consisting of two primary phases. The first phase involves dividing the news graph into multiple clusters. In the second phase, recommendations are generated within each cluster. The findings indicate that this proposed approach yields more precise recommendations while also reducing computational demands.