Infeasibility of Sequitur-Based Motif for Mouse Dynamics in Digital Forensics

Document Type

Conference Proceeding

Source of Publication

12th International Symposium on Digital Forensics and Security, ISDFS 2024

Publication Date

1-1-2024

Abstract

Utilizing and deploying behavioral biometric modalities (BBM), specifically mouse dynamics for user attribution in a digital investigation, has seen a rapid upsurge. However, as asserted in a recent study, the current reliability threshold of BBM falls short of the required standard for forensic attributes. This poor reliability can be attributed, in part, to the low signal-to-noise ratio in a typical behavioral dataset. This study proposed a context-free signature identification and extraction technique for BBM to extract a unique mouse dynamics signature suitable for a forensic process. A Re-Pair Grammar induction approach, which identifies and extracts unique Grammar sequences, was used to achieve this proposition. The grammar generation leverages symbolic aggregate approximation techniques to generate behavioral string subsequences from the mouse dataset. The Re-Pair approach was then used to develop a user attribution mechanism, which can be deployed for digital forensic analysis. The outcome of the implementation of the proposition, however, shows a poor performance relative to existing studies, hence its infeasibility as a benchmark for forensic science. However, it shows promising potential to reveal the inherent noise in mouse dynamics data, which can provide further insight into digital forensic science. This result further extends the literature on establishing digital forensic science, a significant requirement for any forensic discipline.

ISBN

9798350330366

Publisher

IEEE

Disciplines

Computer Sciences

Keywords

behavioral biometrics, Behavioral Motif, digital forensic science, Mouse dynamics signature, Symbolic Aggregate Approximation, user attribution

Scopus ID

85194079093

Indexed in Scopus

yes

Open Access

no

Link to Full Text

Share

COinS