Abstract—In this paper, an intensive study is performed on
the dataset that contains people’s activity records of computer
usage, and a multi-feature based user identification algorithm is
developed. It is shown that features generated from user’s
program/process start up history and website access records are
extremely effective to differentiate people from each other. The
proposed algorithm exploits this fact for identifying user by
matching user’s current behavior features with the history
feature library. The experimental results show that the
proposed multi-feature based algorithm can achieve 91.2% user
identification accuracy.
Index Terms—User identification, computer using behavior,
feature matching, feature combination.
Tongda Zhang is with the Electrical Engineering Department, Stanford
University, Stanford, CA 94305 USA (e-mail: tdzhang@stanford.edu).
Xiao Sun is with National Engineering Laboratory for E-Commerce
Technology, Tsinghua University, Beijing, China; DNSLAB, China Internet
Network Information Center, Beijing 100190, China (e-mail:
sunx11@mails.tsinghua.edu.cn).
Yueting Chai is with the National Engineering Laboratory for
E-Commerce Technology, Tsinghua University, Beijing, China (e-mail:
chaiyt@mail.tsinghua.edu.cn).
Hamid Aghajan is with the Department of Electrical Engineering,
Stanford University, and the Department of Telecommunication and
Informatics in Gent University, Belgium (e-mail: hamid@icdsc.org).
Cite: Tongda Zhang, Xiao Sun, Yueting Chai, and Hamid Aghajan, "Human Computer Interaction Activity Based User Identification," International Journal of Machine Learning and Computing vol.4, no. 4, pp. 354-358, 2014.