Sethemo: thematic segmentation-based ontology

In this paper, we suggest our automatic system SeThemO (Thematic Segmentation-based Ontology) which allows, through the integration of domain ontology, topic borders detection in a specific textual document. We have implemented it on basis of the Conceptual Energy of Auto Associative Memory [Boudouma 13a]. Built with a nested architecture, our system includes seven main modules that we will fully develop in the rest of this article. We also present the performance achieved by the thematic segmentation process SeThemO. We compare it with the results obtained by some known approaches, according to the assessment protocol that we will explain later in this work.