A Flexible Querying Approach Based on Outranking and Classification

* Final gross prices may vary according to local VAT.

Get Access

Abstract

This article presents an approach to database flexible queries which relies on the notion of outranking and a classification process. Instead of being compared pairwise like in Pareto-order-based approaches, tuples are compared to acceptability profiles that are associated with predefined classes. According to their satisfaction of user preferences, tuples are assigned to classes with a certain degree. This approach proposes a classification algorithm with a linear complexity and makes it possible to aggregate non-commensurable preferences.