TY - GEN
T1 - Using a Dynamic Model to Predict Popularity of News
AU - Wang, Xiaomeng
AU - Fang, Binxing
AU - Zhang, Hongli
N1 - Publisher Copyright:
© 2019 IEEE.
PY - 2019/4
Y1 - 2019/4
N2 - Nowadays, with the rapid development of smartphones and wireless networks, real-time news spreads faster because people can use mobile clients to browse images, videos, and audio content. There are several messages in the comment area below the news body where people share their opinion. Some online news items are widely viewed, while the majority draw little attention. Therefore, it is well worth achieving trustworthy predictions of how popular new content may become by calculating the number of users who view, forward, and/or comment on the content. Existing models for popularity prediction are mostly based on independent dissemination of information without competition, and many of them assume that the popularity of hot events does not influence unrelated topics. In this paper, taking Tencent News as a case study, we observe the existence of a popularity migration, whereby, especially when a hot topic happens, people sometimes leave comments on irrelevant topics. The main contribution of this article is in solving the problem that there are few or no works for popularity prediction based on the topic migration effect. We find that the migration popularity of news is well reflected by the competitive strength. We propose a model based on the reinforced Poisson process with the introduction of a competitive matrix.
AB - Nowadays, with the rapid development of smartphones and wireless networks, real-time news spreads faster because people can use mobile clients to browse images, videos, and audio content. There are several messages in the comment area below the news body where people share their opinion. Some online news items are widely viewed, while the majority draw little attention. Therefore, it is well worth achieving trustworthy predictions of how popular new content may become by calculating the number of users who view, forward, and/or comment on the content. Existing models for popularity prediction are mostly based on independent dissemination of information without competition, and many of them assume that the popularity of hot events does not influence unrelated topics. In this paper, taking Tencent News as a case study, we observe the existence of a popularity migration, whereby, especially when a hot topic happens, people sometimes leave comments on irrelevant topics. The main contribution of this article is in solving the problem that there are few or no works for popularity prediction based on the topic migration effect. We find that the migration popularity of news is well reflected by the competitive strength. We propose a model based on the reinforced Poisson process with the introduction of a competitive matrix.
KW - big data
KW - information diffusion
KW - popularity
KW - social networks
UR - https://www.scopus.com/pages/publications/85087607015
M3 - 会议稿件
AN - SCOPUS:85087607015
T3 - INFOCOM 2019 - IEEE Conference on Computer Communications Workshops, INFOCOM WKSHPS 2019
BT - INFOCOM 2019 - IEEE Conference on Computer Communications Workshops, INFOCOM WKSHPS 2019
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2019 INFOCOM IEEE Conference on Computer Communications Workshops, INFOCOM WKSHPS 2019
Y2 - 29 April 2019 through 2 May 2019
ER -