Home // eTELEMED 2015, The Seventh International Conference on eHealth, Telemedicine, and Social Medicine // View article

Diabetes Lifestyle Support with Improved Glycemia Prediction Algorithm

Authors:
Péter Gyuk
Tamás Lőrincz
Rebaz A. H. Karim
István Vassányi

Keywords: Glucose-level tracking; eHealth; Genetic algorithm; Glucose-Insulin system; Glucose absorption; Diabetes mellitus; Outpatient care

Abstract:
This paper proposes a combined model to predict the blood glucose level of people with diabetes. Our method consists of two efficient models found in literature and takes nutrition, applied insulin, and initial glucose level into account during the calculations. An extension has been made to these models using various model training methods. Our aim is to help diabetics calculate the insulin need with this efficient algorithm later implemented in a user-friendly software. The tests, that are based on real data, show a significant improvement in the results if model training methods such as Genetic Algorithm (GA) is used. On the other hand, the numbers reveal the weaknesses of our method, which has to be fixed in the future. During an all-day validation, the prediction error was smaller than 3 mmol/l in 83% of the cases while using GA. Compared to other tests found in literature our model seems to be a good start in predicting glycemia, but needs further improvements.

Pages: 95 to 100

Copyright: Copyright (c) IARIA, 2015

Publication date: February 22, 2015

Published in: conference

ISSN: 2308-4359

ISBN: 978-1-61208-384-1

Location: Lisbon, Portugal

Dates: from February 22, 2015 to February 27, 2015