Short Communication, Jceit Vol: 10 Issue: 1
Continuous Glucose Monitoring Prediction
Akhila Sabbineni*
Department of Microbiology, Andhra University, Vishakhapatnam, India
*Corresponding Author:
Akhila Sabbineni
Department of Microbiology
Andhra University, Vishakhapatnam, India
E-mail: akhilasabbineni@gmail.com
Received: December 22, 2020 Accepted: January 07, 2021 Published: January 14, 2021
Citation: Sabbineni A, (2021) Continuous Glucose Monitoring Prediction J Comput Eng Inf Technol 10:1 DOI: 10.37532/jceit.2021.10(1).247
Abstract
Diabetes is one of the deadliest diseases in the world and affects nearly 10 percent of the global adult population. Fortunately, powerful new technologies allow for a consistent and reliable treatment plan for people with diabetes. One major development is a system called continuous blood glucose monitoring (CGM). In this review we look at three different continuous meal detection methods that were developed given CGM data from patients with diabetes. From this analysis an initial meal prediction algorithm was also developed utilizing these methods.
Keywords: Continuous glucose monitoring; Auto regression; kalman filter; Recurrent neural networks.
Keywords
Continuous glucose monitoring; auto regression; kalman filter; recurrent neural networks.
Introduction
Continuous aldohexose observance mechanically tracks blood sugar levels, additionally referred to as glucose, throughout the day and night. You’ll see your aldohexose level anytime at a look. You’ll additionally review however your aldohexose changes over a number of hours or days to examine trends. Seeing aldohexose levels in real time will assist you build a lot of knowledgeable selections throughout the day concerning a way to balance your food, physical activity, and medicines.
A CGM works through a little device inserted beneath your skin, sometimes on your belly or arm. The device measures your opening aldohexose level that is that the aldohexose found within the fluid between the cells. The device tests aldohexose each jiffy. A transmitter wirelessly sends the knowledge to a monitor.
The monitor is also a part of associate degree hypoglycemic agent pump or a separate device that you may carry during a pocket or purse. Some CGMs send data on to a smartphone or pill. Many models are obtainable and are listed within the Yankee polygenic disease Association’s product guide.
It is vital to find meal-intake in type-1 diabetic patients. Meal intake will have an effect on aldohexose level of the body and it will even cause harmful situations like hyperglycemia.
Fortunately, with the advancement of technology, there are completely different sensors and devices that allow you monitor aldohexose levels. However, several of those devices keep company with their own limitations and inaccuracies. Thus, recursive detection may be a great tool. This project focuses on three such formulas for on-line meal detection associate degreed includes a suggestion for an initial algorithm still. we tend to concentrate on victimisation associate degree Auto-regression primarily based model, Kalman-Filter primarily based approach, associate degreed an LSTM-RNN primarily based approach.
We take off by initial syncing the given CGM statistic (corresponding to a patient) with the bolus ground truth. we tend to then move to the event, representation, and implementation of the three algorithms. we tend to then report the train and take a look at accuracies. when that we tend to proceed to supply execution time analysis and on to the event it's vital to find meal-intake in type-1 diabetic patients. Meal intake will have an effect on aldohexose level of the body and it will even cause harmful situations like hyperglycaemia.
Fortunately, with the advancement of technology, there ar completely different sensors and devices that allow you monitor aldohexose levels. However, several of those devices keep company with their own limitations and inaccuracies. Thus, recursive detection may be a great tool. This project focuses on three such formulas for on-line meal detection associate degreed includes a suggestion for an initial algorithm still. We tend to concentrate on victimisation associate degree Auto-regression primarily based model, Kalman-Filter primarily based approach, associate degreed an LSTM-RNN primarily based approach.
We take off by initial syncing the given CGM statistic (corresponding to a patient) with the bolus ground truth. We tend to then move to the event, representation, and implementation of the three algorithms. We tend to then report the train and take a look at accuracies. When that we tend to proceed to supply execution time analysis and on to the event.
