Skip to Main Content
Recently, mining negative association rules has received some attention and proved to be useful. To the best of our knowledge, three typical forms for negative association rules and three corresponding mining methods have been proposed. However, the existing forms are not general enough, and can not represent some special cases in the real world. In this paper, an extended form for negative association rules is proposed, and a corresponding mining algorithm is presented. The proposed mining algorithm is performed on two datasets. Experimental results show that the algorithm is efficient on simple and sparse datasets when minimum support is high to some degree, and it overcomes some limitations of the previous mining methods. The proposed form will extend related applications of negative association rules to a broader range.
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on (Volume:3 )
Date of Conference: 18-21 Aug. 2005