A Fast Method to Detect Hot Topic from BBS
Xu Hui-jie, Cai Wan-dong, Guirong Chen · 2013
Due to the information of Bulletin Board System (BBS)is numerous, detecting hot topic from BBS is a challenging task.In this paper, we propose a novel method for fast detection of hot topic.By analyzing features of hot topic, we define the hotness indicator to filter out the non-hot threads.Then we fit the changeful curves of threads followed time and remove the threads whose vitality is on the decline.Finally, we give an algorithm that can merge the remaining threads and get hot topics.Results prove that the proposed method is of good validity and feasibility.