PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Segmenting multi-attribute sequences using dynamic Bayesian networks
Robert Gwadera, Janne Toivola and Jaakko Hollmen
In: Seventh IEEE International Conference on Data Mining Workshops (ICDMW 2007), 28 Oct - 31 Oct 2007, Omaha, Nebraska, USA.

Abstract

Discovering dependencies between attributes in multiattribute event sequences (multi-sequences), also known as synchronized multi-stream sequences, is an important problem in many domains, including monitoring systems and molecular biology. Many real-life multi-sequences have a segmental structure, with segments of differing complexities of attribute dependencies, which reflects a changing nature of the dependencies over time and space. In this paper we propose a new approach for discovering dependencies in multi-sequences which considers a possible segmental nature of such dependencies and tries to describe the multisequences in probabilistic terms using Dynamic Bayesian Networks (DBN). To accurately quantify such changing dependencies, we segment the multi-sequence by fitting an optimal DBN for each segment. We use the Bayesian Information Criterion (BIC) to select an optimal DBN structure and the number of segments of the multi-sequence.

PDF - PASCAL Members only - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Conference or Workshop Item (Paper)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Computational, Information-Theoretic Learning with Statistics
Learning/Statistics & Optimisation
Theory & Algorithms
ID Code:3610
Deposited By:Jaakko Hollmen
Deposited On:13 February 2008