Hemoglobin A1C and Fructosamine Correlation: Assessing Glycemic Control in Diabetic Hypertensive Patients

Abstract

This paper introduces a new computer-based approach called the MAG-T framework that will allow finding blood sugar and blood pressure risks and correct the problem of not identifying all of them that is present with the traditional metabolic tests. It utilizes a model that considers the relationship of words in order to clarify the relationship between Hemoglobin A1c (HbA1c), Fructosamine, and blood pressure changes. The approach extracts information on time and metabolism by techniques such as TF-IDF, CBOW, and N-Skip Gram (SKG). It applies random up-sampling to manage the issue of smaller data on atypical high-risk cases. The process is more efficient, and the model is stabilized using Variance Threshold Feature Selection (VTFS) as well as Predictor Testing (SPT). The framework applies a set of nine machine learning classifiers such as SVM-RBF, the Random Forest (RF), and ANN-LM, as well as majority voting to make final risk predictions. The tests made using the MAG-T framework outperform the available models with an accuracy of 97.1, F-Measure of 0.96, and AUC of 0.999. It uses Temporal Kinetic Mapping (TKM) plus consensus-based sorting to detect concealed metabolic instability that is overlooked by peer-reviewed HbA1c tests. The reliability of the model is validated using validation tests that test the ability of the model to cope with measurement errors. The MAG-T framework is a valuable instrument to personal endocrinology and heart risk stratification because of its sensitivity to high-risk types.