Digestive casebase mining based on possibility theory and linear unidimensional scaling
Résumé
In this paper, we present a very general and powerful approach that enables us to mine easily depending on the concept of similarity any casebase consisting of a large number of objects (cases) containing heterogeneous, imperfect and missing data by organizing and gathering these objects into meaningful groups in such a way that efficient analysis and retrieval of information could be easily achieved. Our method is based essentially on possibility theory and on the linear unidimensional scaling representation and is applied on a real digestive database.