| Summary: | Introduction: Diabetic retinopathy (DR) is a complication of diabetes mellitus (DM) and a leading cause of vision loss among adults in Malaysia. The severity of DR is influenced by risk factors such as the duration of DM, diabetes control, and systemic comorbidities. There are five stages of DR severity: no DR, mild non-proliferative DR (NPDR), moderate NPDR, severe NPDR, and proliferative DR (PDR). Predicting DR severity can assist medical professionals in prioritizing patient care. Methods: This study aimed to predict DR severity based on patient risk factors using a dataset from the Ophthalmology Clinic, UiTM. Machine learning (ML) algorithms, including XGBoost, Random Forest, Decision Tree, Support Vector Machine, and Gradient Boosting Tree, were utilized. These models were tested with random state values of 10, 20, 30, 40, and 50. Model performance was evaluated using accuracy, precision, recall, F1-Score, and Cohen’s Kappa. Results: Among the models, Random Forest with a random state value of 50 demonstrated the best performance: an accuracy of 90.61%, precision of 91.02%, recall of 90.61%, F1-Score of 90.28%, and Cohen’s Kappa of 88.26%. This model was subsequently used to develop a web-based prediction system. Conclusion: The prediction system facilitates early identification of DR severity in diabetic patients. The findings of this research can significantly contribute to future studies and improve patient care in this field. © 2024 Universiti Putra Malaysia Press. All rights reserved.
|
|---|