Using Coupled Hidden Markov Models to Model Suspect Interactions in Digital Forensic Analysis

Authors:
Nathan Brewer;Nianjun Liu;Olivier De Vel;Terry Caelli
Affiliations:
The Australia National University, Australia/ National ICT Australia (NICTA), Australia;The Australia National University, Australia/ National ICT Australia (NICTA), Australia;Defence Science and Technology Organisation (DSTO), Australia;The Australia National University, Australia/ National ICT Australia (NICTA), Australia
Venue:
AIDM '06 Proceedings of the International Workshop on on Integrating AI and Data Mining
Year:
2006

Citing 0
Cited 2

Semantic retrieval of events from indoor surveillance video databases

Pattern Recognition Letters
HMM-Based Defect Localization in Wire Ropes --- A New Approach to Unusual Subsequence Recognition

Proceedings of the 31st DAGM Symposium on Pattern Recognition

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper expands the use of Hidden Markov Models in Digital Forensics by using Coupled Hidden Markov Models to investigate interactions between multiple suspects in forensic cases. This paper compares the output of a coupled hidden Markov model to a similarly trained single-chain hidden Markov models and to expert knowledge in a simulated digital forensic case with two suspects. In the situation modelled, there was a notable improvement in the accuracy of the coupled model compared to single chain models. This demonstrates that there is some form of interaction between suspects in this digital forensic case, and that this interaction can be effectively modelled by a coupled hidden Markov model.