Rough Association Rule Mining in Text Documents for Acquiring Web User Information Needs

  • Authors:
  • Yuefeng Li;Ning Zhong

  • Affiliations:
  • Queensland University of Technology, Australia;Maebashi Institute of Technology, Japan

  • Venue:
  • WI '06 Proceedings of the 2006 IEEE/WIC/ACM International Conference on Web Intelligence
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

It is a big challenge to apply data mining techniques for effective Web information gathering because of duplications and ambiguities of data values (e.g., terms). To provide an effective solution to this challenge, this paper first explains the relationship between association rules and rough set based decision rules. It proves that a decision pattern is a kind of closed pattern. It also presents a novel concept of rough association rules in order to improve the effectiveness of association rule mining. The premise of a rough association rule consists of a set of terms and a frequency distribution of terms. The distinct advantage of rough association rules is that they contain more specific information than normal association rules. It is also feasible to update rough association rules dynamically to produce effective results.