Data Mining Using Granular Computing: Fast Algorithms for Finding Association Rules
T. Lin, Eric K. Louie · Studies in fuzziness and soft computing · 2002
An attribute value of a relation is a meaningful name(common property) of a group of entities (elementary granule). A relational model using such elementary granules as its attribute values is called machine oriented relational model. In such a model, data processing, in particular finding association rules is transformed into granular computing. In this paper, algorithms for finding association rules by granular computing is presented. Analysis and experiments show that the computation is fast and is a promising approach. Experiments show about 15–20 time faster; theoretical analysis indicates that on the counting the support step, which is the major step, it is at least 32 (wordsize) time faster.