LINEAR REGRESSION AND FUZZY LOGIC IN THE ANALYSIS OF FACTORS INFLUENCING GLYCATED HEMOGLOBIN IN CHILDREN AND ADOLESCENTS WITH TYPE 1 DIABETES MELLITUS

Abstract

This article presents a study on the analysis of factors that influence glycated hemoglobin (HbA1c) levels in children and adolescents with type 1 diabetes mellitus (DM1). After an initial characterization of the dataset, linear regression and fuzzy logic were used to model and interpret these factors, with an additional focus on nutritional status and insulin delivery method. The study covers a sample of 80 patients between 4 and 19 years of age and is based on data collected at a medical outpatient clinic in a city in the interior of São Paulo. The results indicate a significant difference in HbA1c levels between insulin delivery methods (SICI and MDI), but not between different nutritional statuses. The linear regression model points to the time of diagnosis and total cholesterol as the most significant factors for the model. Thus, the fuzzy logic complemented the analysis of the study, allowing a three-dimensional visual representation of the variables and showing that the two variables significantly affect the levels of HbA1c.