A fuzzy-ontology-oriented case-based reasoning framework for semantic diabetes diagnosis

Area of Science:

• Artificial Intelligence
• Medical Informatics
• Knowledge Representation

Background:

• Case-based reasoning (CBR) solves problems using past experiences.
• Integrating formal ontologies enhances CBR, especially in healthcare.
• Handling vague medical knowledge is crucial for intelligent systems.

Purpose of the Study:

• To propose a fuzzy ontology-based case-based reasoning (CBR) framework.
• To develop a fuzzy semantic retrieval algorithm for improved case matching.
• To enhance the semantic intelligence of CBR systems for medical applications.

Main Methods:

• Developed a fuzzy ontology-based CBR framework using OWL2.
• Implemented a fuzzy semantic retrieval algorithm to handle diverse feature types.
• Tested the framework on a diabetes diagnosis problem with 60 real cases.

Main Results:

• The system achieved 97.67% accuracy in diabetes diagnosis.
• The fuzzy ontology comprised 63 concepts and 2640 instances.
• The framework effectively handles complex medical queries and vague terms.

Conclusions:

• Integrated CBR systems demonstrate improved performance.
• Fuzzy ontologies and advanced retrieval algorithms enhance CBR accuracy.
• The proposed framework offers a robust solution for intelligent medical diagnosis.