Efficient parallel algorithms for mining associations

Mahesh V. Joshi, Eui Hong Sam Han, George Karypis, Vipin Kumar

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Scopus citations

Abstract

The problem of mining hidden associations present in the large amounts of data has seen widespread applications in many practical domains such as customer-oriented planning and marketing, telecommunication network monitoring, and analyzing data from scientific experiments. The combinatorial complexity of the problem and phenomenal growth in the sizes of available datasets motivate the need for efficient and scalable parallel algorithms. The design of such algorithms is challenging. This chapter presents an evolutionary and comparative review of many existing representative serial and parallel algorithms for discovering two kinds of associations. The first part of the chapter is devoted to the non-sequential associations, which utilize the relationships between events that happen together. The second part is devoted to the more general and potentially more useful sequential associations, which utilize the temporal or sequential relationships between events. It is shown that many existing algorithms actually belong to a few categories which are decided by the broader design strategies. Overall the aim of the chapter is to provide a comprehensive account of the challenges and issues involved in effective parallel formulations of algorithms for discovering associations, and how various existing algorithms try to handle them.

Original languageEnglish (US)
Title of host publicationLarge-Scale Parallel Data Mining
EditorsMohammed J. Zaki, Ching-Tien Ho
PublisherSpringer Verlag
Pages83-126
Number of pages44
ISBN (Print)3540671943, 9783540671947
DOIs
StatePublished - 2002
Event5th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 1999 - San Diego, United States
Duration: Aug 15 1999Aug 15 1999

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1759
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other5th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 1999
CountryUnited States
CitySan Diego
Period8/15/998/15/99

Bibliographical note

Publisher Copyright:
© Springer-Verlag Berlin Heidelberg 2000.

Fingerprint Dive into the research topics of 'Efficient parallel algorithms for mining associations'. Together they form a unique fingerprint.

Cite this