WCSE 2016
ISBN: 978-981-11-0008-6 DOI: 10.18178/wcse.2016.06.075

Automata Bases Associative Classification (AAC) for Data Mining

Mohammad Abrar, Alex Tze Hiang Sim, Jee Mei Hee

Abstract— The study on the use of association rules for the purpose of classification gave rise to a new field known as Associative Classification (AC). The process used to generate association rules is exponential by nature; thus in AC, researcher focused on the reduction of redundant rules via rules pruning and rules ranking techniques. The removal of rules however could negatively affect accuracy. In this paper, we radically store most of the rules in a condensed form utilizing automata. The automata offsets critical need for rules pruning and ranking. Our new structure is used for classification. Experimental results show that the accuracy of our automata based technique is significantly improved compare to the existing state-of-the-art algorithms which includes J48, AODE, BayesNet and FT etc. The analysis also shows that our automata based associative classification technique is efficient by means of computational time and space utilization.

Index Terms— associative classification, automata, association rules, data mining, machine learning.

Mohammad Abrar, Alex Tze Hiang Sim
Faculty of Computing, Universiti Teknologi Malaysia, MALAYSIA
Jee Mei Hee
Faculty of Education, Universiti Teknologi Malaysia,MALAYSIA

[Download]


Cite: Mohammad Abrar, Alex Tze Hiang Sim, Jee Mei Hee, "Automata Bases Associative Classification (AAC) for Data Mining," Proceedings of 2016 6th International Workshop on Computer Science and Engineering, pp. 449-452, Tokyo, 17-19 June, 2016.