Granule mining oriented data warehousing model for representations of multidimensional association rules

Abstract

To promise the quality of multidimensional association mining in real applications is a challenging research issue. The challenging issue is how to represent multidimensional association rules efficiently because of the complicated correlation between attributes. Multi-tier granule mining is one initiative in solving this challenge. This paper presents a granule mining oriented data warehousing model. It can divide attributes into tiers and discover granules for each tier from large multidimensional databases. In addition, it uses association mappings to generate association rules for describing the correlation between tiers. Experiments for the proposed model and the testing results are prosecuted.