Efficient mining of association rules based on gravitational search algorithm
Association rules mining are one of the most used tools to discover
relationships among attributes in a database. A lot of algorithms
have been introduced for discovering these rules. These algorithms
have to mine association rules in two stages separately. Most of
them mine occurrence rules which are easily predictable by the
users. Therefore, this paper discusses the application of
gravitational search algorithm for discovering interesting
association rules. This evolutionary algorithm is based on the
Newtonian gravity and the laws of motion. Furthermore, contrary
to the previous methods, the proposed method in this study is able
to mine the best association rules without generating frequent
itemsets and is independent of the minimum support and
confidence values. The results of applying this method in
comparison with the method of mining association rules based
upon the particle swarm optimization show that our method is
successful.
Keywords: Association Rules, Gravitational Search Algorithm, Swarm intelligence
Download Full-Text