An Intelligent Knowledge Based System for Diagnosis and Treatment of Diabetes

Background/Objectives: Diabetes poses a significant health challenge in Ethiopia due to factors such as lifestyle choices, a lack of medical professionals, and limited laboratory resources. Knowledge-based system (KBS) are increasingly being explored to assist healthcare providers in diagnosing and managing diseases, including diabetes. Rule-based and case-based reasoning are two primary approaches to developing these knowledge-based systems. The major goal of this research is to create an intelligent KBS for diabetes diagnosis and therapy utilizing data mining approaches. Methods: This study uses a design science research approach to develop a smart, knowledge-based system for diagnosing and managing diabetes, combining rule-based and case-based reasoning. Data were collected from Hiwot Fana Specialized Hospital at Haramaya University and Harari Regular Hospital, with the dataset undergoing preprocessing to address errors, noise, outliers, duplicates, missing values, and normalization. Rule-based reasoning was implemented using the Prolog programming language, while the case-based reasoning system was built with JCOLLIBIR studio using its CBR main cycle. For predictive and descriptive modeling, the J48 decision tree, PART, and JRip algorithms were evaluated for rule-based reasoning, while K-means and Farthest First were evaluated for case-based reasoning. Findings: The JRip rule induction algorithm produced superior results for the rule-based system, while the K-means clustering algorithm performed best for case-based reasoning. These outputs were used to construct the rule base and automatically generate the case base, respectively. The integrated system achieved an overall accuracy of 95.23%, with a precision of 100% and a recall of 89.7%. Additionally, 93.2% of users passed the user acceptance test, demonstrating the system's practical utility. Novelty and Applications: This study presents a novel integration of rule-based and case-based reasoning using a rule-dominant approach to create an intelligent knowledge-based system for diabetes diagnosis and management. The system has shown high accuracy and user satisfaction, making it a viable tool for aiding healthcare professionals in Ethiopia, especially in resource-limited settings. Keywords: Rule Based Reasoning, Case Based Reasoning, Diagnosis and Treatment of Diabetes, Intelligent Knowledge-Based System