A Knowledge Representation Approch For Ethiopian Traditional Medicine

Abstract

Traditional herbal medicines are heavily used in third world nations to treat different types of diseases and Ethiopia is no exception. Even though usage of western medicine is prevalent, Ethiopians still rely more on traditional medicine. Garlic, ginger and others used to treat commonly recurring illnesses such as cold. Traditional medicine is very complex and wide spread In Ethiopia. Different ethnic groups uses different variety of herbs to cure diseases. However, there is a lack of computable data that contain plant species that are used as herbal medicine. This creates a struggle for researchers to gather and compile information about traditional medicine that can be used for different research purposes. The purpose of this thesis is to model an otology based KR prototype by employing an ontology structure for Ethiopian Traditional Herbal Medicine. This allows domain experts who engaged in this specific area to capture and record semantically rich knowledge. Since it uses the OWL Semantic Web standard frame work, it can be integrated with in a linked-data platform to be shared over the web. The flexibility and scalability of ontology schema can greatly enhances the richness and expressiveness of the KR ontology model. Besides representing the domain area, the computing community in Ethiopia can take this research work as a supportive material in order to introduce ontologies as a capable KR modeling tool. The open source ontology editor and management platform called Prot??g?? is used to build the prototype. SPARQL query scripts are generated to extract a specific sets of information in the prototype knowledge base. Since SPARQL is the standard query language of the Semantic Web applications, it?�?s very important to utilize it in the context of this thesis. Another very easy querying mechanism is the DL Query which is integrated with in the Prot??g?? platform. This particular querying method allows a domain expert to formulate a query script and extract a specific piece of information from the ontology based KR model. Visualization is also important as it depicts the overall or specific architecture of the system in a glance. Third party application called AllegroGraph is used to generate the SPARQL query as well as the visualization that accompanies it. The proposed prototype ontology is qualitatively evaluated by domain experts and it showed promising results in representing the domain knowledge. Eighty nine percent acceptability rate is gained after the evaluation session is done.