文章基本信息

标题：Classification Rule and Exception Mining Using Nature Inspired Algorithms
本地全文：下载
作者：Amarnath Pathak ; Jyoti Vashistha
期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2015
卷号：6
期号：3
页码：3023-3030
出版社：TechScience Publications
摘要：Classification is an important data mining task which facilitates list of decision rules that helps us to predict class of an unseen instance. Various traditional techniques like Decision tress, Neural Networks, SVMs have been used in past for rule mining. Nature Inspired Algorithms (NIAs) are class of algorithms that mimic natural processes and are capable of mining comprehensible and accurate rules. It is interesting to investigate Nature Inspired Algorithms (NIAs), exclusively GA and ACO, in context of rule mining. Classification model usually represents obvious information in form of decision rules and an unseen instance is liable to be misclassified if the model created using any of the above techniques do not account for exceptions present in the dataset. Instances having low support count and deviating from obvious behavior are termed as exceptions and they are less likely to be discovered using the usual rule discovery measures that account for generality of the discovered knowledge. In this paper we have investigated use of NIAs in rule mining and exception mining and we have also suggested possible modification in existing cAntMinerpb algorithm for mining exceptions.
关键词：Nature Inspired Algorithms; Genetic;Algorithm (GA); Ant Colony Optimization (ACO);Exceptions.