Semantic Knowledge Fusion in Healthcare: A Hybrid Approach for Connected Medicine

Abstract

In a context where connected medicine requires increasingly explainable, accurate, and responsive systems, this paper presents an applied experimental research focusing on the development and evaluation of a hybrid intelligent assistant for healthcare data fusion. The study is based on the parallel combination of two data paradigms: classical tabular structures and their ontological equivalent. Using an intelligent assistant, we simultaneously query a medical dataset on diabetes in tabular form and the same dataset translated into an OWL ontology that can be queried using SPARQL. The aim is to demonstrate that the synchronised combination of these two models not only provides a more complete response but also one that is better contextualised and clinically exploitable. The research follows an experimental methodology, involving the implementation, testing, and comparative evaluation of both models on 300 questions classified by increasing complexity (simple, complex, and very complex). The results reveal a relevance rate above 99%, with an average response time suited to medical use. This work highlights the potential of hybrid architectures in connected health and paves the way for new decision-making assistants that fully exploit the semantic richness of medical knowledge.