Abstract—A sporadic rule is an association rule which has low support but high confidence. In general, sporadic rules are of rare occurrence but high value in many cases. Of the two types of perfectly and imperfectly sporadic rules, imperfectly sporadic rules are more difficult to mine since they consist of individual items with high support whereas the support of combinations of these items is low. The problem of mining imperfectly sporadic rules has not been completely solved till now. Thus, the paper describes an absolute answer to the question by proposing a problem of mining imperfectly sporadic rules with two thresholds and developing a MCISI (mining closed imperfectly sporadic itemsets) algorithm to find imperfectly sporadic itemsets with two thresholds. The development of MCISI algorithm is based on a closed itemset lattice, therefore efficiency of the algorithm can be improved through reduction of search space and removal of redundant imperfectly sporadic rules with two thresholds. We also point out that mining imperfectly sporadic rules could be considered as a special case of mining imperfectly sporadic rules with two thresholds, and imperfectly sporadic rules with two thresholds are of rare occurrence comparing with imperfectly sporadic rules.
Index Terms—Rare Association Rule; Imperfectly Sporadic Rule; Imperfectly Sporadic Rule with Two Thresholds.
Cu Thu Thuy is a teacher at Economic Information System - Academy of Finance, Ha Noi, Viet Nam (e-mail: cuthuthuy@hvtc.edu.vn).
Do Van Thanh is a researcher and scientific manager. He has been working for National Center for Scio-Economic Information and Forecast, Ha Noi, Viet Nam (e-mail: Thanhdv_db@mpi. gov.vn).
Cite: Cu Thu Thuy and Do Van Thanh, "Mining Imperfectly Sporadic Rules with Two Thresholds," International
Journal of Computer Theory and Engineering vol. 2, no. 5, pp. 718-723, 2010.
Copyright © 2008-2024. International Association of Computer Science and Information Technology. All rights reserved.
