Real-time Monitoring and Early Detection of Diabetes with Bioactive and Biological Impedance Sensors using Hybrid Machine Learning Algorithm

Abstract

Diabetes is an ongoing infection that influences a large number of individuals overall and can prompt serious unexpected issues whenever left untreated. Early identification of diabetes can altogether diminish the risk of intricacies and work on significant results. Lately, the utilization of wearable technology has arisen as a promising device for illness identification and checking. Smartwatches furnished with bioactive sensors can give ceaseless, painless observing of body vitals, making them ideal for diabetes screening. This study proposes a framework that uses patient information for preparing a hybrid AI model to distinguish the presence of diabetes. The framework consolidates body vitals estimated utilizing a smartwatch with a bioactive sensor to get exact and nonstop information on the wearer's wellbeing status. The mixture model coordinates both profound learning and conventional AI calculations to accomplish predominant precision in identifying diabetes. The framework gathers information on different body vitals, for example, pulse, circulatory strain, and skin conductance, which are known to be firmly connected with diabetes. The gathered information is pre-handled and afterward used to prepare the hybrid model. The profound learning calculation is utilized to remove significant level highlights from the crude information, while the conventional AI calculation is utilized to arrange the information into diabetic or non-diabetic classifications. The cross breed model is intended to work on the accuracy of diabetes location by integrating the qualities of both profound learning and conventional AI.