A Tweet-Centric Algorithm for News Ranking

Ranking news is helpful because of the information explosion which overloads readers. It is also a challenging task since news is now published by both news portals and microblogging platforms in real time. Most traditional news ranking algorithms consider two factors: media focus and user attention independently. While in the paper, we propose a news ranking framework to combine the two factors together. A better ranking algorithm is obtained via the following two parts: (1) the Influence Method (IM) to rank news and (2) the News Flow Graph Method (NFGM) to locate the most influential source for duplicated news from multiple news sources. We present four strategies to evaluate user attention. Experiments show that decay strategy based on Ebbinghaus forgetting curve is the best one. To the best of our knowledge, our paper is the first attempt to utilize microblogging data for news ranking.