A clustering based approach for query relaxation in evidential databases

Queries posed by a user over a database do not always return the desired responses. It may sometimes result an empty set of answers especially when data are pervaded with uncertainty and imprecision. Thus, to address this problem, we propose an approach for relaxing a failing query in the context of evidential databases. The uncertainty in such databases is expressed within the belief function theory. The key idea of our approach is to use a machine learning method more precisely the belief K-modes clustering technique to relax the failing queries by modifying the constraints in order to provide successful alternatives which may be of interest to the user.