A Collaborative Filtering Based Ranking Algorithm for Classifying and Ranking NEWS TOPICS Using Factors of Social Media

Searching topic and tracking the headlines on Topic Detection and Tracking (TDT) makes the users more convenient to see what is happening in the real-world through internet. Due to the large quantity of news articles, it is not possible to see all the topics. It makes the necessity of topic ranking in terms of time and importance. Topic ranking is based on frequency of occurrence of the topic in media and the amount of attention paid by the users. Both these factors are time dependent. So it is necessary to include the effect of time. However, inconsistency always exists between these two factors. In this paper, an automatic online news topic ranking algorithm has been proposed. The analysis between Media Focus (MF) and User Attention (UA) has been carried out in the proposed algorithm in terms of inconsistency. Here, UI defines the strength of the community who discusses the topic. Overlapping Topic Clusters (TCs) is found by Hybrid Fuzzy Clustering (HFC) approach. Artificial Bee Colony Optimization (ABCO) is performed to calculate the node weight. It is necessary to personalize the SociRank to present different topics for different users. Collaborative Filtering based Ranking Algorithm (CFRA) is adopted for ranking. The proposed CFRA based SociRank improves the quality and variety of automatically identified news topics. The same has been proven by experimental results.