News Recommendation using Social Network Data

Abstract

Many social networks exist today- for example, Facebook and twitter. They have played a major role in our day to day news consumption. These sources have a huge potential to provide important news related information. These include user-generated data as well as numerous links to online articles. However, these are not completely useful for all types of users without a proper filtering process. The news is filled with noise as well as irrelevant media that may not be considered valuable. In addition, it is vital to prioritize this data post the filtering process due to the possibility of information overload. This can be attained by ranking the information by making use of three major factors. The first factor is the media focus (MF) of any topic in the news which describes its temporal prevalence in the news. The second factor is the user attention (UA) which indicates its temporal prevalence in social network media. The last factor is known as user interaction (UI) which depicts the various interactions amongst all the users who are currently discussing this topic. This paper presents a solution that uses an unsupervised framework to extract and rank news topics from social media and news sources according the above mentioned factors.