Elimination Algorithm of Redundant Association Rules Based on Domain Knowledge

Abstract

Many association rule mining algorithms have been developed to extract interesting patterns from large databases. However, a large amount of knowledge explicitly represented in domain knowledge has not been used to reduce the number of association rules. A significant number of known associations are unnecessarily extracted by association rule mining algorithms. The result is the generation of hundreds or thousands of non-interesting association rules. This paper presents an algorithm named DKARM, which takes into account not only database itself, but also related domain knowledge, so as to eliminate extraction of known associations in domain knowledge. Experiments show this algorithm can reasonably eliminate redundant rules, and effectively reduce the number of rules.