Granular Association Analysis

Yun Shen, Trevor P. Martin

Abstract: The topic of fuzzy information granulation was first proposed by Zadeh in 1979. There has been a fast increasing interest in this topic and its applications since in many situations, when a problem involves incomplete, vague or uncertain information, it is difficult to come to a crisp solution and requires a granular solution which has a very significant impact on the design and implementation of intelligent systems. Care is needed when information systems use fuzzy sets to model graded membership in granules - the fuzzy sets are not disjunctive possibility distributions, but must be interpreted conjunctively. This distinction is clarified theoretically in this paper and it is shown how an extended mass assignment framework can be used to extract relations between granules. These relations are association rules and are useful when integrating multiple information sources. Experimental results on film genres, commercial database and terrorism incident databases carried out in the SQuAD project are demonstrated and discussed in this paper.