Composite Ontology-Based Medical Diagnosis Decision Support System Framework

Abstract

Current medical decision support systems have evolved from the automation of medical decision routines to improving the quality of health care services. Knowledge-based systems, compared to conventional data-driven techniques, are promising to support medical decision making. However, knowledge acquisition is usually a bottleneck in the process of developing such systemsOne possibility for acquiring medical knowledge, particularly tacit knowledge, is to use data or cases in both syntactic and semantic ways. Case-based Reasoning (CBR) methodology provides a practical way of problem solving with recalled knowledge memory of solved cases. To reduce the difficulty of knowledge acquisition, this paper proposes a design of the system framework that utilizes the simplified medical knowledge:disease-symptom ontology for prediagnosis, given patients symptoms and signs as input. In the first stage, simple pattern matching is used to gather candidate diseases in diagnosis. Following that, case-based reasoning is used to refine diagnostic decision. The case base is structured with ontological knowledge model. The case retrieval process is based on semantic similarity. The diagnostic system uses a composite knowledge base, and will allow automated diagnosis recommendation. The system framework also aims at facilitating semantic explanations to the solution derived.