Abstract:
Clinical diagnosis and monitoring are gradually shifting into the homes of patients. These are now easily attained because of increased efforts geared towards creation of more approximate medical diagnostic reasoning algorithms (MDRAs). Researchers have crafted different medical algorithms for developing healthcare delivery systems, with most of these algorithms being built on the knowledge clinicians have learnt through study and experience during diagnostic procedures. Some of these models are based on statistical, mathematical, fuzzy and rule-based techniques. Despite the differences in their underlying approaches, they are all oriented towards a MDRA with higher precision. In this research, we developed an enhanced MDRA that is particularly addressing the limitations of the reasoning functions of an MDRA called Select and Test (ST) algorithm. The logical inference making process of ST is being limited by its use of simple logical constructs and some applications of mathematical methods. Therefore, the semantic-based MDRA framework this paper presents builds on the ST reasoning structures, aided by using the semantic web concept. We then model the knowledgebase using an ontological approach, design and implement a coordinated rule system for effective reasoning and uses semantic web-based rule/reasoning engines for rule implementation and inference making respectively. This enhanced framework adds a monitoring agent that autonomously improves both its knowledge base and to actualize its monitoring task. We use our enhanced MDRA as a test bed for breast cancer diagnosis and designed a set of metrics for comparing the result of our improved ST algorithm with the existing ST algorithm.
Keywords: Semantic Web, Inference making, Ontology, Rule set and Diagnosis
