Système de gestion de la médecine traditionnelle dans une plateforme web social et sémantique : une approche basée sur une ontologie visuelle

Abstract : Traditional medicine (TM) is seen as a complementary health system to modern medicine (MM). This TM has the advantage of being easier to access, cheaper and more adapted to local specificities. However, African TM suffers from the lack of sharing and preservation of medical knowledge : the majority of practitioners speak local languages. They use non-codified languages, and many of them keep their knowledge secret. To facilitate knowledge preservation and communication between practitioners of TM (TMP) we aim at designing SysMEDTRAD, a system based on a domain ontology of TM and a visual ontology integrating an iconic language. This research work brings an innovative solution for practitioners of TM, not only to allow them to contribute continuously to the modernization and the explicit improvement of this medicine, but also to equip them with a tool to practice their healing art. More precisely, this work has focused on the construction of a domain ontology referred to as ontoCONCEPT-Term entirely terminological. It also includes an iconic language started on identifying medicinal plants. Traditional recipes and remedies have a diagrammatic representation. This language will have to be extended while following the princple drawn by a visual ontology approach, ontoICONE. Via this visual ontology, we can trenscand the barriers related to illiteracy, orality and the many local languages to the number of 1127 in West Africa, where this community web knowledge sharing project takes place. ontoICONE is partly terminological and iconic. This terminological part is common with ontoCONCEPT-Term. Those two entities ontoICONE and ontoCONCEPT-Term form ontoMEDTRAD. To build our ontology, we took inspiration from ontology construction methodologies, especially Diligent, NeOn and OntoForInfoScience. We have combined multiple sources of information and knowledge on TM, including repeated meetings with TMP in different regions, published works in biosciences (TM, ethnobotanics) covering the West African sub-region, websites, books, observation in botanic gardens, national structures of TM (PNPMT), NGOs (Prometra), associations and federations of TMP (Côte d’Ivoire, Sénégal). We identified three main use cases : identifying the disease from the patient’s symptoms and signs ; prescribing the remedy (determining the abstract and material recipes : its ingredients and how to prepare it) ; preparing the remedy and administering it to the patient. The richness of TM, its tacit and taboo nature, and the complexity of the visual representation, have led us to adopt a progressive construction approach based on semiotic rules namely visual similarity, semantic association and convention. In medicinal resources, plants represent more than 60% of TM’s treatments. We have therefore chosen to initiate iconization with the Plant concept by designing an iconic language from the morphological properties of the plant based on botanical sources. We have collected a dataset of 22 plants and 31 antimalarial remedies, and we have verified that our iconic language allows us to identify these 22 plants, using an automatic learning and classification tool (Weka). This iconic language was presented to TMP during a first validation step. The next step will focus on the definitive and global validation of ontoMEDTRAD, and mostly the adoption of a real visual ontology (ontoICONE). It also integrates the deployment of ontoMEDTRAD (ontoICONE) into a semantic wiki platform, such as Virtuoso, in order to have SysMEDTRAD being operational. In perspectives, the T-Box and A-Box of a formal ontology will be completed by G-Box, a graphic part including the icons and their structure (ontoMEDTRAD). Align the visual ontology with existing ones (TM and MM) and give a regional and worldwide reference stature to this domain ontology of African TM will be not left out.
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/tel-01842116
Contributor : Appoh Kouame <>
Submitted on : Wednesday, March 27, 2019 - 3:26:05 PM
Last modification on : Tuesday, April 9, 2019 - 5:01:42 PM

File

ThKAppoh2.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01842116, version 9

Citation

Appoh Kouame. Système de gestion de la médecine traditionnelle dans une plateforme web social et sémantique : une approche basée sur une ontologie visuelle. Intelligence artificielle [cs.AI]. Université Gaston Berger de Saint-Louis (Sénégal), 2018. Français. ⟨tel-01842116v9⟩

Share

Metrics

Record views

14

Files downloads

63