A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Data Science and Artificial Intelligence

Thesis (M. Sc.) - Asian Institute of Technology, 2024

Diabetes care calls both medical knowledge and the interpretation of several laboratory test findings. But in hospital systems, medical records often just show lab findings, which require additional evaluation by physicians. A system with an organized dashboard with analysis results, tracking, and notification functions will improve the accuracy and time of medical management. This paper creates a diabetes clinical decision support system. The CDSS analyzed and displayed lab results in a dashboard which allows quick interpretation from the user. Based on the clinical observations of doctors, clinical data collection offers many levels of indicators. The features were chosen using the feature significance selection procedure and previous work. The models were selected from the best of related work. The results of the experiment show that an applied dataset with the suggested feature selections gives an accuracy of 89% via using a Support Vector Machine (SVM) model in comparison to Random Forest, Decision Tree, K-Nearest neighbors, and Logistic regression classifiers.The dashboard was designed to show analyzed data with color-code which represents the control risk. The design of the system was co-created from experiments, interviews, and evaluations with doctors. All functions combine into the system with input, prediction, and display features using the Microsoft service environment mainly in Microsoft Power Apps.