Mining negative rules: a literature review focusing on performance

Abstract

Mining of frequent patterns and association rules is a Data Mining task that aims to determine consistent relationships among elements in a transaction database. Algorithms that consider the absence of elements perform the generation of so-called negative rules which result in associations of great interest for some applications, enabling it to obtain extra knowledge in comparison to the positive case. This type of association presents a problem regarding the increased amount of generated rules which demands adequate computational resources. This study presents a systematic review with the aim of grouping the concepts of the main contemporary works on this topic, in order to assist the development of future works in this subject.