Towards a hybrid method of construction of a normalized domain ontology used by machine teaching PERO2

Abstract

The challenges facing learning of reasoning require a general refounding of knowledge modeling expressed by the development and integration of intelligence and reasoning techniques in the processing of this knowledge. In this paper, we present a part of the research work of PERO2 project. This intelligent system is a machine teaching dedicated to the learning of reasoning and problem-solving of exercises in physical science. Our research is focused on representing the knowledge base of PERO2 by integrating a semantic layer based upon a domain ontology, capable of adding some intelligence and enriching the reasoning in our system. In order to build this ontology, we propose a hybrid method based on two main phases: (*) design phase of our domain ontology called “OntoPhyScEx”. (**) Semantic validation phase of this ontology.