New algorithms for efficient mining of association rules

Li Shen, Hong Shen, Ling Cheng

Research output: Contribution to journalArticle

29 Citations (Scopus)

Abstract

Discovery of association rules is an important data mining task. Several algorithms have been proposed to solve this problem. Most of them require repeated passes over the database, which incurs huge I/O overhead and high synchronization expense in parallel cases. There are a few algorithms trying to reduce these costs. But they contain weaknesses such as often requiring high pre-processing cost to get a vertical database layout, containing much redundant computation in parallel cases, and so on. We propose new association mining algorithms to overcome the above drawbacks, through minimizing the I/O cost and effectively controlling the computation cost. Experiments on well-known synthetic data show that our algorithms consistently outperform a priori, one of the best algorithms for association mining, by factors ranging from 2 to 4 in most cases. Also, our algorithms are very easy to be parallelized, and we present a parallelization for them based on a shared-nothing architecture. We observe that our parallelization develops the parallelism more sufficiently than two of the best existing parallel algorithms.

Original languageEnglish (US)
Pages (from-to)251-268
Number of pages18
JournalInformation Sciences
Volume118
Issue number1
DOIs
StatePublished - Sep 1999
Externally publishedYes

Fingerprint

Association rules
Association Rules
Mining
Costs
Parallelization
Synthetic Data
Parallel algorithms
Parallel Algorithms
Parallelism
Data mining
Preprocessing
Layout
Data Mining
Synchronization
Vertical
Processing
Experiment
Experiments

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • Information Systems
  • Information Systems and Management
  • Statistics, Probability and Uncertainty
  • Electrical and Electronic Engineering
  • Statistics and Probability

Cite this

New algorithms for efficient mining of association rules. / Shen, Li; Shen, Hong; Cheng, Ling.

In: Information Sciences, Vol. 118, No. 1, 09.1999, p. 251-268.

Research output: Contribution to journalArticle

Shen, Li ; Shen, Hong ; Cheng, Ling. / New algorithms for efficient mining of association rules. In: Information Sciences. 1999 ; Vol. 118, No. 1. pp. 251-268.
@article{d9af02f5aab24f348a0cf50d38330dc0,
title = "New algorithms for efficient mining of association rules",
abstract = "Discovery of association rules is an important data mining task. Several algorithms have been proposed to solve this problem. Most of them require repeated passes over the database, which incurs huge I/O overhead and high synchronization expense in parallel cases. There are a few algorithms trying to reduce these costs. But they contain weaknesses such as often requiring high pre-processing cost to get a vertical database layout, containing much redundant computation in parallel cases, and so on. We propose new association mining algorithms to overcome the above drawbacks, through minimizing the I/O cost and effectively controlling the computation cost. Experiments on well-known synthetic data show that our algorithms consistently outperform a priori, one of the best algorithms for association mining, by factors ranging from 2 to 4 in most cases. Also, our algorithms are very easy to be parallelized, and we present a parallelization for them based on a shared-nothing architecture. We observe that our parallelization develops the parallelism more sufficiently than two of the best existing parallel algorithms.",
author = "Li Shen and Hong Shen and Ling Cheng",
year = "1999",
month = "9",
doi = "10.1016/S0020-0255(99)00035-3",
language = "English (US)",
volume = "118",
pages = "251--268",
journal = "Information Sciences",
issn = "0020-0255",
publisher = "Elsevier Inc.",
number = "1",

}

TY - JOUR

T1 - New algorithms for efficient mining of association rules

AU - Shen, Li

AU - Shen, Hong

AU - Cheng, Ling

PY - 1999/9

Y1 - 1999/9

N2 - Discovery of association rules is an important data mining task. Several algorithms have been proposed to solve this problem. Most of them require repeated passes over the database, which incurs huge I/O overhead and high synchronization expense in parallel cases. There are a few algorithms trying to reduce these costs. But they contain weaknesses such as often requiring high pre-processing cost to get a vertical database layout, containing much redundant computation in parallel cases, and so on. We propose new association mining algorithms to overcome the above drawbacks, through minimizing the I/O cost and effectively controlling the computation cost. Experiments on well-known synthetic data show that our algorithms consistently outperform a priori, one of the best algorithms for association mining, by factors ranging from 2 to 4 in most cases. Also, our algorithms are very easy to be parallelized, and we present a parallelization for them based on a shared-nothing architecture. We observe that our parallelization develops the parallelism more sufficiently than two of the best existing parallel algorithms.

AB - Discovery of association rules is an important data mining task. Several algorithms have been proposed to solve this problem. Most of them require repeated passes over the database, which incurs huge I/O overhead and high synchronization expense in parallel cases. There are a few algorithms trying to reduce these costs. But they contain weaknesses such as often requiring high pre-processing cost to get a vertical database layout, containing much redundant computation in parallel cases, and so on. We propose new association mining algorithms to overcome the above drawbacks, through minimizing the I/O cost and effectively controlling the computation cost. Experiments on well-known synthetic data show that our algorithms consistently outperform a priori, one of the best algorithms for association mining, by factors ranging from 2 to 4 in most cases. Also, our algorithms are very easy to be parallelized, and we present a parallelization for them based on a shared-nothing architecture. We observe that our parallelization develops the parallelism more sufficiently than two of the best existing parallel algorithms.

UR - http://www.scopus.com/inward/record.url?scp=0033320738&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033320738&partnerID=8YFLogxK

U2 - 10.1016/S0020-0255(99)00035-3

DO - 10.1016/S0020-0255(99)00035-3

M3 - Article

VL - 118

SP - 251

EP - 268

JO - Information Sciences

JF - Information Sciences

SN - 0020-0255

IS - 1

ER -