Please use this identifier to cite or link to this item:
Title: MaxRI : a method for discovering maximal rare itemsets
Author(s): Darrab, Sadeq
Broneske, DavidLook up in the Integrated Authority File of the German National Library
Saake, GunterLook up in the Integrated Authority File of the German National Library
Issue Date: 2021
Type: Conference object
Language: English
URN: urn:nbn:de:gbv:ma9:1-1981185920-882376
Subjects: Rare itemsets
Representative rare itemsets
Maximal rare itemsets
Abstract: Rare itemset mining got extensive attention due to its high importance in real-life applications. Rare itemset mining methods aim at discovering the whole set of rare itemsets in a dataset. Although current algorithms perform reasonably well in finding interesting rare itemsets, they also reveal a large number of rare itemsets, including redundant ones. As a result, skimming through these massive amounts of (partly redundant) itemsets is a big overhead in many applications. On the other hand, generating a massive number of rare itemsets also compromises the performance of algorithms in terms of time and memory. To address these limitations, we propose an efficient algorithm called maximal rare itemset (MaxRI) to discover maximal rare patterns (long rare itemset). Then, we propose another method RRI (Recover Rare Itemsets from maximal rare itemsets) to retrieve the interesting subset of rare itemsets of a user-given length, k, from the set of maximal rare itemsets. To the best of our knowledge, this is the first paper proposed for rare itemset mining by considering the representative rare patterns without redundant ones. Our experimental results indicate that our proposed methods’ performance is better than the up-to-date algorithms in terms of time and memory consumption.
Open Access: Open access publication
License: (CC BY 4.0) Creative Commons Attribution 4.0(CC BY 4.0) Creative Commons Attribution 4.0
Sponsor/Funder: Transformationsvertrag
Appears in Collections:Fakultät für Informatik (OA)

Files in This Item:
File Description SizeFormat 
Darrab et al._MaxRI_2021.pdfZweitveröffentlichung633.9 kBAdobe PDFThumbnail