Sequential Association Rule Mining Revisited: A Study Directed at Relational Pattern Mining for Multi-morbidity

Vincent-Paulraj, A, Burnside, G ORCID: 0000-0001-7398-1346, Coenen, F ORCID: 0000-0003-1026-6649, Pirmohamed, M ORCID: 0000-0002-7534-7266 and Walker, L ORCID: 0000-0002-3827-4387 (2021) Sequential Association Rule Mining Revisited: A Study Directed at Relational Pattern Mining for Multi-morbidity .

Text paulraj_SGAI2021.pdf - Author Accepted Manuscript Download (302kB) | Preview

Abstract

Sequential rule mining is a well-established data mining technique for binary valued data. Many variations have been proposed, most approaches use the support-confidence-lift framework. Existing approaches make assumptions concerning the definition of what a sequence is. However, this definition is application dependent. In this paper we look at sequential rule mining with respect to multi-morbidity disease prediction which entails a rethink of the definition of what a sequence is, and a consequent rethink of the operation of the support-confidence-lift framework. A novel sequential rule mining algorithm is proposed designed to address the challenge of multi-morbidity disease prediction. The SEquential RElational N-DIsease Pattern (SERENDIP) algorithm.

Item Type:	Conference Item (Unspecified)
Uncontrolled Keywords:	Sequential rule mining, Multi-morbidity disease prediction
Divisions:	Faculty of Health & Life Sciences Faculty of Health & Life Sciences > Inst. Population Health Faculty of Health & Life Sciences > Inst. Systems, Molec & Integrative Biology > Inst. Systems, Molec & Integrative Biology Faculty of Science & Engineering > School of Electrical Engineering, Electronics and Computer Science
Depositing User:	Symplectic Admin
Date Deposited:	12 Oct 2021 10:48
Last Modified:	24 Jan 2026 03:06
DOI:	10.1007/978-3-030-91100-3_20
Related Websites:	Author Publisher
URI:	https://livrepository.liverpool.ac.uk/id/eprint/3140162
Disclaimer:	The University of Liverpool is not responsible for content contained on other websites from links within repository metadata. Please contact us if you notice anything that appears incorrect or inappropriate.