Your Age Is No Secret: Inferring Microbloggers' Ages via Content and Interaction Analysis
Jinxue Zhang, Xia Hu, Yanchao Zhang, Huan Liu
Abstract
Microblogging systems such as Twitter have seen explosive use in public and private sectors. The age information of microbloggers can be very useful for many applications such as viral marketing and social studies/surveys. Current microblogging systems, however, have very sparse age information. In this paper, we present MAIF, a novel framework that explores public content and interaction information in microblogging systems to explore the hidden ages of microbloggers. We thoroughly evaluate the accuracy of MAIF with a real-world dataset with 54,879 Twitter users. Our results show that MAIF can achieve up to 81.38% inference accuracy and outperforms the state of the art by 9.15%. We also discuss some countermeasures to alleviate the possible privacy concerns caused by MAIF.