A Novel Approach for Blog Feeds Recommendation Based on Meta-data

Abstract

As the blogosphere continues to grow, finding good quality blog feeds has been very time consuming and requires much effort. So, recommending blog feeds, which handle topics close to user interests, can be useful. Recently, the number of bloggers who use the subscription services has been increasing. Subscription is a service using protocols like RSS and ATOM, which notify users when new entries are posted on the blogs that the users register for subscription. In this paper, we present an effective and efficient approach to recommending log feeds based on the subscription lists and meta-data of blogs. In order to find blogs that handle topics close to the blogs in subscription lists, we first model the topic of blogs by collecting and expanding tags in the blogs, and we then compare the topic models of blogs for recommendation. Also, the usefulness of blogs is an important factor. For choosing useful blogs, we adopt the update frequency and number of subscribers because useful blogs are the ones in which new entries will be frequently posted and to which many users will subscribe. In order to validate the proposed blog recommendation algorithm, experiments on real blog data have been conducted, and the results of experiments show that our blog feed recommendation can satisfy users who want to subscribe to blog feeds.