A dynamic bibliometric model for identifying online communities

Predictive modelling of online dynamic user-interaction recordings and community identifi cation from such data b ecomes more and more imp ortant w ith th e w idesp read use of online communication tech nologies. D esp ite of th e time-dep endent nature of th e p rob lem, ex isting ap p roach es of community identifi cation are b ased on static or fully ob served netw ork connections. H ere w e p resent a new , dynamic generative model for th e inference of communities from a seq uence of temp oral events p roduced th rough online comp uter-mediated interactions. T h e distinctive feature of our ap p roach is th at it tries to model th e p rocess in a more realistic manner, including an account for p ossib le random temp oral delays b etw een th e intended connections. T h e inference of th ese delays from th e data th en forms an integral p art of our state-clustering meth odology, so th at th e most lik ely communities are found on th e b asis of th e lik ely intended connections ra...

• Type: Journal
• Year: 2008
• Where: DATAMINE
• Authors: Xin Wang, Ata Kabán
• Keywords: Community Identifi Cation, DATAMINE 2008, Etw Een Th, Identifi Cation