An Interestingness Measure and Computation Method of Association Rules Based on Frequent Itemsets Relatedness

Abstract:

Article Preview

To address inadequacy of association rules interestingness measure method currently, we present a novel method to measure interestingness with relatedness among items in frequent itemsets. It firstly computed relatedness between frequent k-itemsets and each subset of frequent 2-itemsets, which is a linear combination of Complementarity Intensity (CI), Substitutability Intensity (SI) and Mutual Interaction (MI). The mean of relatedness of all frequent 2-itemsets subsets was regarded as relatedness of frequent k-itemsets. Finally weighted computation method of association rule interestingness was given according to principle of objective interestingness of association rule is inversely proportional to relatedness of frequent itemsets. The method can not only sort rules, but also analyze actual relationship among all items in frequent 2-itemsets, which is conductive to selection of users on rules.

Info:

Periodical:

Edited by:

Dongye Sun, Wen-Pei Sung and Ran Chen

Pages:

4039-4043

DOI:

10.4028/www.scientific.net/AMM.71-78.4039

Citation:

X. Chen et al., "An Interestingness Measure and Computation Method of Association Rules Based on Frequent Itemsets Relatedness", Applied Mechanics and Materials, Vols. 71-78, pp. 4039-4043, 2011

Online since:

July 2011

Export:

Price:

$35.00

In order to see related information, you need to Login.

In order to see related information, you need to Login.