Indexed by:Journal Article

Date of Publication:2016-08-01

Journal:COGNITIVE COMPUTATION

Included Journals:Scopus、EI、SCIE

Volume:8

Issue:4

Page Number:577-586

ISSN:1866-9956

Key Words:Topic detection; Emotion distribution; Chinese microblog

Abstract:Peoples' perceptions of reality are conditioned on how others see the world. Unfortunately, with the vast amount of information made available through online media, such as microblog sites, it is impossible for people to absorb all information in a timely manner. Therefore, the detection of hot topics on a microblog platform is becoming increasingly important. The present paper proposes a new hot-topic detection and extraction approach based on language and topic models, which analyzes the differences in emotion distribution language models between adjacent time intervals to detect hot topics. According to the contents and repost degree of microblogs, we estimate the importance of each microblog and generate topic models. Experiments conducted on the Sina Microblog show that the proposed approach can detect and extract hot topics effectively and can thus assist the Sina Microblog platform in managing and monitoring hot topics.