Genetic Algorithms for Classification Rule Discovery: Issues and Challenges

Abstract: Classification Rule discovery is one of the major tasks of data mining. Genetic algorithms (GAs) for discovery of classification rules have recently gained increased importance because these are capable of finding global optimal solutions in large and complex search spaces like classification rule discovery. Building a rule based classification model, particularly for large data sets, using Genetic Algorithms becomes a very complex and time-consuming task. This paper enlists some of important issues that influence the efficacy and/or efficiency of GAs in the domain of building classification models. The paper also emphasizes the research direction to make application of GAs more amiable for discovering classification rules.

Keywords: Genetic Algorithm, Data Mining, Classification, Rule Mining