Machine Learning Techniques for Risk Assessment and Diagnosis of Diabetes Mellitus

Abstract

Diabetes represents a group of diseases that are characterized by hyperglycemia resulting from defects in insulin secretion, insulin action, or both. Machine Learning algorithms are increasingly used for risk assessment and diagnosis of various disease, so this paper represents development of machine learning-based model used for risk assessment and diagnosis of diabetes mellitus. For development of this system 300 samples consisting of information about Fasting Plasma Glucose (FPG), Oral Glucose Tolerance Test (OGTT) and blood test called HgbA1c were used. All samples were obtained from several healthcare institutions in Bosnia and Herzegovina, and diagnosis of diabetes mellitus and healthy patients in this dataset were established by medical professionals. Machine learning techniques for risk assessment and diagnosis of diabetes mellitus, were trained with 210 samples. Testing of developed system was performed with 90 samples for validation. The developed system shows 98,89% accuracy while classifying diabetes.