Mining differential dependencies: A subspace clustering approach

S Kwashie, Jixue Liu, Jiuyong Li, Feiyue Ye

Research output: Book chapter/Published conference paperConference paperpeer-review

12 Citations (Scopus)


The discovery of differential dependencies (DDs) is the problem of finding a minimal cover set of DDs that hold in a given relation. This paper proposes a novel subspace-clustering-based approach to mine DDs that exist in a given relation. We study and reveal a link between delta-nClusters and differential functions (DFs). Based on this relationship, we adopt and co-opt techniques for mining delta-nClusters to find the set of candidate antecedent DFs of DDs efficiently, based on a user-specified distance threshold. Furthermore, we define an interestingness measure for DDs to aid the discovery of essential DDs and avoid the mining of an extremely large set. Finally, we demonstrate the scalability and efficiency of our solution through experiments on real-world benchmark datasets.
Original languageEnglish
Title of host publicationDatabases Theory And Applications, ADC 2014
EditorsHua Wang, Mohamed A. Sharaf
Place of PublicationBerlin, Germany
PublisherSpringer-Verlag Berlin Heidelberg
Number of pages12
ISBN (Electronic)9783319086071
ISBN (Print)9783319086088
Publication statusPublished - 2014
EventAustralasian Database Conference: ADC 2014 - University of Queensland, Brisbane, Australia
Duration: 14 Jul 201416 Jul 2014


ConferenceAustralasian Database Conference


Dive into the research topics of 'Mining differential dependencies: A subspace clustering approach'. Together they form a unique fingerprint.

Cite this