Hierarchy as a new data type for qualitative variables

摘要

The concept of hierarchy has being explored by the computer science communities during last few decades. Relatively simple hierarchical structures found extensive use in such diverse areas as data modeling, information retrieval, knowledge representation and processing, natural language, pattern recognition, and so on. Recent investigations in information retrieval and data integration have emphasized the use of ontologies and semantic similarity functions as a mechanism for comparing objects that can be retrieved or integrated across heterogeneous repositories. Hierarchies being a simpler, albeit very useful, version of ontologies, can perfectly contribute to model solutions of these problems. Present paper aims to illustrate above thesis by discussing a simple method of information retrieval that uses a hierarchical qualitative data organization. Its main goal is to retrieve objects from any database that are just close to a desired item and control the retrieval process up to a given error, called herein confusion. For doing this, we define a semantic dissimilarity (confusion) between objects to be retrieved as well as introduce a calculus of predicates based on the confusion function.