Granular Association Rules for Multiple Taxonomies: A Mass Assignment Approach

摘要: The use of hierarchical taxonomies to organise information (or sets objects) is a common approach for the semantic web and elsewhere, based on progressively finer granulations objects. In many cases, seemingly crisp granulation disguises fact that categories are loosely defined concepts better modelled by allowing graded membership. A related problem arises when different used, with structures, as integration process may also lead fuzzy categories. Care needed systems model membership in - not disjunctive possibility distributions, but must be interpreted conjunctively. We clarify this distinction show how an extended mass assignment framework can used extract relations between These association rules useful integrating multiple sources categorised according hierarchies. Our do suffer from problems associated cardinalities. Experimental results discovering film databases terrorism incident demonstrated.