An Incremental Rare Association Rule Mining Approach with a Life Cycle Tree Structure Considering Time-Sensitive Data
Hu Kerui,Qiu Lemiao,Zhang Shuyou,Wang Zili,Fang Naiyu
DOI: https://doi.org/10.1007/s10489-022-03978-3
IF: 5.3
2022-01-01
Applied Intelligence
Abstract:One of the association rule mining techniques, rare association rule mining (RARM), is a method for finding association rules with low support but high confidence from a database. In some fields, a large amount of time-sensitive data, where old data may lose their influence over time, are constantly generated. A combination of valid and invalid data and a change in database cause low efficiency in extracting effective rare association rules. To solve this issue, we propose an efficient incremental RARM method with a novel tree structure inspired by the concept of the tree life cycle. The tree structure is a compressed data structure that can store transactional and temporal information as tree paths and node attributes, respectively, through a single database scan. Upon receiving new data, the tree can be adjusted and updated through a fast restructuring strategy without rebuilding it from scratch, which reduces time and memory consumption. To avoid interference from invalid data, the tree adopts a damped window model to evaluate the value of each transaction and removes invalid transactions before the rule mining process. Therefore, the rare association rules extracted from the life cycle tree structure in the time-sensitive data environment are more common and reliable. Comprehensive experiments on real and synthetic datasets are conducted, and the results show that the proposed method performs better in terms of runtime, memory usage, and scalability than the state-of-the-art methods.