Automatic Mining of Numerical Classification Rules with Parliamentary Optimization Algorithm

Advances in Electrical and Computer Engineering

KIZILOLUK, S¹ 

关键词: classification algorithms;    computational intelligence;    data mining;    heuristic algorithms;    optimization;   

DOI  :  10.4316/AECE.2015.04003

学科分类：计算机科学（综合）

来源: Universitatea "Stefan cel Mare" din Suceava

【 摘 要 】

In recent years, classification rules mining has been one of the most important data mining tasks. In this study, one of the newest social-based metaheuristic methods, Parliamentary Optimization Algorithm (POA), is firstly used for automatically mining of comprehensible and accurate classification rules within datasets which have numerical attributes. Four different numerical datasets have been selected from UCI data warehouse and classification rules of high quality have been obtained. Furthermore, the results obtained from designed POA have been compared with the results obtained from four different popular classification rules mining algorithms used in WEKA. Although POA is very new and no applications in complex data mining problems have been performed, the results seem promising. The used objective function is very flexible and many different objectives can easily be added to. The intervals of the numerical attributes in the rules have been automatically found without any a priori process, as done in other classification rules mining algorithms, which causes the modification of datasets.

【 授权许可】

Unknown