From Social User Activities to People Affiliation

摘要: This study addresses the problem of inferring users' employment affiliation information from social activities. It is motivated by applications which need to monitoring and analyzing activities employees a given company, especially their tracks related work business. definitely helps better understand needs opinions towards certain business area, so that account sales targeting these customers in company can adjust strategies accordingly. Specifically, this task we are snapshot network some labeled users who company. Our goal identify more same We formulate as classifying nodes over graph, develop Supervised Label Propagation model. naturally incorporates rich set features for activities, models networking effect label propagation, learns feature weights labels propagated right users. To validate its effectiveness, show our case studies on identifying "China Telecom" Unicom" Sina Weibo. The experimental results method significantly outperforms compared baseline ones.