Mining temporal association rules with incremental standing for segment progressive filter

Abstract

Association rule mining is a popular data mining technique which dredges up valuable relationships among different items in a dataset. A variant called temporal association rule mining finds relationship between items with respect to particular time periods. Databases are frequently updated; therefore temporal association rules that we discover should be corresponding to the updates in databases. Most of the existing data mining techniques however do not cover revising associate rules from the latest updates in the dataset. Some form of incremental mining technique is also needed to embrace the fresh elements that are updated continuously in the transaction database. In this paper we propose a technique that modifies the frequent patterns in pace with changes to the database over time. An Incremental Standing method for Segment Progressive Filter (ISPF) is proposed. ISPF algorithm is used for supporting the temporal association rule mining in transaction database with different exhibition periods. Our algorithm is optimized such that scanning of database is minimized. Scan reduction technique is applied here to generate all candidate k-item sets to form 2-candidate item sets directly. Working of the proposed algorithm is tested and illustrated with examples and a case study respectively. © 2011 Springer-Verlag.