Diagnosis and recommender system for diabetes patient using decision tree / Nurul Aida Mohd Zamary

Diabetes mellitus, particularly known as type 2 diabetes, is a significant public health concern in Malaysia. The increasing prevalence of this chronic disease necessitates the development of effective diagnosis and recommender systems to reduce its impact on public healthcare. This project aims to develop a decision-making support model for diabetes diagnosis and treatment recommendation using the decision tree algorithm. The objectives include studying the requirements of the decision tree in the diagnosis and recommendation system, developing a prototype for the system, and evaluating the accuracy of the decision tree algorithm. The phase of this project is divided into data preprocessing, implementation of the decision tree algorithm, and evaluation of the algorithm and prototype. The decision tree algorithm demonstrated good performance in classifying diabetes mellitus and providing treatment recommendations with accuracy of 98.15% and 98.03%, respectively. To evaluate the model, the model accuracy, precision, recall, F1- score, and confusion matrix were used. The decision tree model in this project is also compared to Naive Bayes and AdaBoost. The decision tree model shows good performance for this decision-making support model. The development of this decisionmaking support model holds significant implications for the early diagnosis and effective management of diabetes mellitus. By providing accurate classification and treatment recommendations, this model has the potential to improve patient outcomes and reduce the time for decision-making in the healthcare industry. The project successfully achieved its objectives by analyzing the literature, developing the decision tree algorithm, and evaluating the accuracy of the model. The findings underscore the potential of decisionmaking support models for improving diabetes diagnosis and treatment recommendation systems.