Data Integration, Approximate Categorisation and Fuzzy Associations
Trevor Martin, Yun Cheng Shen · Bristol Research (University of Bristol) · 2009
The use of hierarchical taxonomies to organise information (or sets of objects) is essential to the semantic web and is also fundamental to many aspects of web 2.0. In most cases, the seemingly crisp granulation of a taxonomy disguises the fact that categories are based on loosely defined concepts which are better modelled by allowing graded membership. Fuzzy categories may also arise when integrating information from multiple sources which do not conform to precisely the same taxonomy definitions. Knowledge of relations between categories can be summarised by association rules. In this paper, we outline a new method to calculate fuzzy confidences for association rules between fuzzy categories from different hierarchies. We illustrate with examples drawn from a system that integrates information from web-based sources.