A framework to mine high-level emerging patterns by attribute-oriented induction

Muyeba, Maybin K.; Khan, Muhammad S.; Warnars, Spits; Keane, John

Abstract

This paper presents a framework to mine summary emerging patterns in contrast to the familiar low-level patterns. Generally, growth rate based on low-level data and simple supports are used to measure emerging patterns (EP) from one dataset to another. This consequently leads to numerous EPs because of the large numbers of items. We propose an approach that uses high-level data: high-level data captures the data semantics of a collection of attributes values by using taxonomies, and always has larger support than low-level data. We apply a well known algorithm, attribute-oriented induction (AOI), that generalises attributes using taxonomies and investigate properties of the rule sets obtained by generalisation algorithms.

• Presentation Conference Type: Conference Paper (published)
• Conference Name: Intelligent Data Engineering and Automated Learning -- IDEAL 2011, 12th International Conference
• Start Date: Sep 7, 2011
• End Date: Sep 9, 2011
• Publication Date: 2011
• Deposit Date: Apr 8, 2025
• Print ISSN: 0302-9743
• Publisher: Springer Verlag
• Peer Reviewed: Peer Reviewed
• Series Number: 6936
• Book Title: Intelligent Data Engineering and Automated Learning -- IDEAL 2011
• ISBN: 978-3-642-23877-2
• DOI: https://doi.org/10.1007/978-3-642-23878-9_21