Using Semantic Graphs in Clustering Process : Enhance Information Level

In this paper, we particularly focused our attention on how to enhance expressivity of ontologies when used as organized space values in a catalogue request process. Using the Wishbone engine [Method for a data-driven adjustement of queries over a database and system for implementing the method], we describe our new contribution by using a semantic graph instead of a tree for the solution set clustering step. We perform a dynamic transformation in order to enhance the intension meaning, and implement a linear programming clustering algorithm. We also expose a rating system in order to evaluate our clustering process quality.