Full Citation

Title: Mining Association Rules with Weighted Items

Citation Type: Dissertation/Thesis

Publication Year: 1998

ISBN:

ISSN:

DOI:

NSFID:

PMCID:

PMID:

Abstract: Mining of association rules has been found useful in many applications. Many previous works focused on the basket (binary) association rules, which is in the form of "The transactions show that there are many customers who purchase product A will purchase the product B." All the products are treated uniformly, and all the rules are mined based on the counts of products. However, in the social science research, the analysts may want to mine the rules based on the importance of the products, items or attributes. For example, total income attribute is more interesting than the height of a person in a household. Based on this, we generalize this to the case where the items are given weights to reflect the importance to the users. As the downward closure property of the support measure in the mining of association rules no longer exists, previous algorithms cannot be applied. In this thesis, we make use of a metric, called support bounds, in the mining of binary and qualitative (fuzzy) association rules. Furthermore, we introduce the simple sample method and the data maintenance method, based on a statistical approach, to mine the rules. Experiments show that by using the weights, we can prune these items without any interest at an earlier step, and hence saving time in mining of these association rules.

User Submitted?: No

Authors: Cai, Cun Hing

Institution: Chinese University of Hong Kong

Department: Department of Computer Science and Engineering

Advisor: Ada W. C. Fu

Degree: Master of Philosophy

Publisher Location: Hong Kong

Pages:

Data Collections: IPUMS USA

Topics: Fertility and Mortality

Countries: