Article ID: | iaor20084047 |
Country: | United Kingdom |
Volume: | 38 |
Issue: | 2 |
Start Page Number: | 203 |
End Page Number: | 228 |
Publication Date: | Feb 2007 |
Journal: | Cybernetics and Systems |
Authors: | Hu Yi-Chung |
Keywords: | fuzzy sets |
Fuzzy sequential patterns are discovered by finding intertransaction fuzzy patterns among data items at a single level. In this paper, a fuzzy data mining method for finding fuzzy sequential patterns at multiple levels of abstraction is developed. Actually, a new method is proposed to mine multiple level fuzzy sequential patterns, using fuzzy partition by a simple fuzzy grid, among data items at a concept hierarchy. The proposed method is composed of two phases: one to find frequent level crossing fuzzy l-sequences, and the other to generate multiple level fuzzy sequential patterns by analyzing the temporal relation between those frequent fuzzy l-sequences. A numerical example along with a mining process is used to illustrate the usefulness of the proposed method.