Comparison of Heuristics for Optimization of Association Rules

Article type: Research Article

Authors: Alsolami, Fawaz^a | Amin, Talha^a | Moshkov, Mikhail^a | Zielosko, Beata^{b; *} | Żabiński, Krzysztof^b

Affiliations: [a] Computer, Electrical and Mathematical Sciences and Engineering Division, King Abdullah University of Science and Technology, Thuwal 23955-6900, Saudi Arabia. fawaz.alsolami@kaust.edu.sa, talha.amin@kaust.edu.sa, mikhail.moshkov@kaust.edu.sa | [b] Institute of Computer Science, University of Silesia in Katowice, 39 Będzińska St., 41-200 Sosnowiec, Poland. beata.zielosko@us.edu.pl, krzysztof.kamil.zabinski@gmail.com

Correspondence: [*] Address for correspondence: Institute of Computer Science, University of Silesia in Katowice, 39 Będzińska St., 41-200 Sosnowiec, Poland.

Abstract: In this paper, seven greedy heuristics for construction of association rules are compared from the point of view of the length and coverage of constructed rules. The obtained rules are compared also with optimal ones constructed by dynamic programming algorithms. The average relative difference between length of rules constructed by the best heuristic and minimum length of rules is at most 4%. The same situation is with coverage.

Keywords: Greedy heuristics, association rules, decision rules, dynamic programming, rough sets

DOI: 10.3233/FI-2019-1791

Journal: Fundamenta Informaticae, vol. 166, no. 1, pp. 1-14, 2019