Mining association rules on time-sensitive data

Abstract

Real-world data is continuously produced and collected quickly; hence, mining association rules are used to fnd
relationships between item sets. One of these combined rule
mining techniques is Rare Association Rule Mining (RARM),
a method that extracts rare, combined rules with low support
but high confdence from the database. Additionally, a large amount of time-sensitive data is created in many felds, and combining valid and outdated data leads to low efciency in extracting combined rules. This thesis explores a mining method that increases rare, combined rules on time-sensitive data to extract common and rare patterns in the database.
The primary approach of the research proposes a new tree structure called ILC-tree, inspired by the LC-tree structure published in 2022. Specifcally, our approach includes a new tree structure and a fast reconstruction method. The proposed approach reduces execution time and memory consumption
by approximately 2 times compared to the old tree structure.