Real-time Detection and Sorting of News on Microblogging Platforms

Real-time Detection and Sorting of News on Microblogging Platforms Wenting Tu David W. Cheung Nikos Mamoulis Min Yang Ziyu Lu Department of Computer Science The University of Hong Kong Pokfulam, Hong Kong {wttu,dcheung,nikos,myang,zylu}@cs.hku.hk Abstract Due to the increasing popularity of microblog- ging platforms (e.g., Twitter), detecting real- time news from microblogs (e.g., tweets) has recently drawn a lot of attention. Most of the previous work on this subject detect news by analyzing propagation patterns of microblogs. This approach has two limitations: (i) many non-news microblogs (e.g. marketing activi- ties) have propagation patterns similar to news microblogs and therefore they can be false- ly reported as news; (ii) using propagation patterns to identify news involves a time de- lay until the pattern is formed, therefore news are not detected in real time. We propose an alternative approach, which, motivated by the necessity of real-time detection of news, does not rely on propagation of posts. More- over, we propose a real-time sorting strategy that orders the detected news microblogs us- ing a translational approach. An experimen- tal evaluation on a large-scale microblogging dataset demonstrates the effectiveness of our approach. 1 Introduction Microblogging platforms (e.g., Twitter or SinaWei- bo) have become very popular and their role as news media has been recognized. As people actively talk about what is happening, microblogs are the place where the first-hand news appear. Actually, over 85% of the leading topics on Twitter are news by nature (H. P et al, 2010). Most of the recent works on news detection from microblogs rely on using temporal patterns of propa- gation (G. L. et al, 2010; R. G. and K. Lerman, 2010; J. L and J. Yang, 2011; F. Z. et al, 2012). As an in- stance, (F. Z. et al, 2012) assumes that bursty topics in microblogs correspond to events that have attract- ed the most online attention. To find such events, this work uses a model to detect busty topics, as- suming that a global event is likely to follow a time- dependent global topic distribution. Although detec- tion methods relying on the propagation characteris- tics of microblogs are based on reasonable assump- tions, they have certain limitations. First, some mi- croblogs not related to news have very similar prop- agation characteristics as news microblogs. For ex- ample, a microblog with a promotion or a gift may follow a similar propagation pattern as a popular news event. Second, propagation behavior can on- ly be analyzed after a microblog has been posted for a certain amount of time. Previous work on news de- tection based on propagation knowledge cannot per- form real-time detection, since propagation knowl- edge can only be obtained a period of time after mi- croblogs are published. Some works explicitly men- tion that trying to detect microblogs using propaga- tion knowledge in a short time reduces the effective- ness. For example, in (G. L. et al, 2010), experi- ments on Twitter data show that using 1-day prop- agation knowledge can mainly detect topics related to …

Content truncated to 500 words