Multi-modal Fusion for Online Users Behavior Recognition

Presented by Sanjida Tumpa as a part of the 2020 Serene-risc Workshop on The State of Canadian Cybersecurity Conference: Human-Centric Cybersecurity

About the presentation

Identifying individuals based on their digital footprints has gain popularity in the field of ubiquitous biometric research. Users disclose unique behavioral trails on the Online Social Networking (OSN) platforms through their shared thoughts and interactions within their respective networks of influence. Cybersecurity researchers focus on analyzing those trails for the purpose of ensuring online security. This research investigates whether users’ self-written contents shared on Twitter can be used for their biometric identification. A Term Frequency–Inverse Document Frequency (TF-IDF) based vocabulary set is generated for every user to serve this purpose. Furthermore, the combination of users’ writing behavior with their acquaintance networks and interests is incorporated in this research using different fusion techniques. The performances of score-level and rank-level fusion algorithms to combine different types of online social behavioral traits are compared. The research concludes that users’ writing behaviors possess uniqueness that allows to obtain higher accuracy than that of other OSN traits, and that score level fusion algorithms result in better accuracy than other fusion methods. The proposed system is able to achieve the recognition rates of 97.72% using the weighted sum rule algorithm at rank-1 and 99.86% at rank-10 after cross-validation. Such system can be used for the purpose of continuous authentication in the cyberworld, which is the focus of further research. 

About the speaker

Sanjida Nasreen Tumpa is currently pursuing her M.Sc. in Computer Science at the University of Calgary, Canada, under the supervision of Prof. Marina L. Gavrilova. She received a B.Sc. degree in Computer Science and Engineering (CSE) from the Military Institute of Science and Technology (MIST), Bangladesh in 2014 and an M.Sc. degree in CSE from Bangladesh University of Engineering and Technology (BUET), Bangladesh in 2019. She served the Department of CSE, MIST, Bangladesh as a faculty member from 2015 to 2019. She has published over 15 conference papers, in addition to two book chapters. Her research interests include social network analysis, natural language processing, and machine learning.