Comparative Analysis of Decision Tree and Neural Network Models for Diabetes Type Classification Based on Clinical Symptoms

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Accurately diagnosing and classifying diabetes types based on clinical symptoms remains a significant challenge in healthcare, particularly in designing effective treatment and control strategies. This study applies decision tree and neural network classification models to categorize diabetes types using symptom-based data obtained from medical records at All Saints’ Hospital, Owerri, Imo State, Nigeria (2024). While the decision tree model employs a rule-based approach to identify key predictors, the neural network leverages its capacity to model complex, non-linear relationships. The values of the average squared classification error (ASCE) and the correct classification rate (CCR) indicate that the neural network model performs better than the decision tree model. The neural network model achieved a CCR of 31 and an ASCE of 32.6, while the decision tree model recorded a CCR of 27.66 and an ASCE of 88.33. Thus, the neural network model is identified as the superior approach for symptom based diabetes classification in this context.