Interpretation of Long-Term Clinical Diabetes Data and Prediction of Glycemic Control Based on Data Mining Techniques

  • E. I. Georga
  • V. C. Protopappas
  • C. Bellos
  • D. Makriyiannis
  • D. I. Fotiadis
Part of the IFMBE Proceedings book series (IFMBE, volume 42)

Abstract

Diabetes is a lifelong disease requiring the active and continuous participation of both diabetic individuals and physicians in order to be well managed and controlled. Technological advancements in the field of diabetes have the potential to improve both self-monitoring and clinical management of the disease. In this study, we present a machine-learning approach to the prediction of daily glucose time series as well as of hypoglycemic events, which relies on support vector and Gaussian processes for regression. For this purpose, medical and lifestyle information, recorded by patients on a daily basis, is exploited. In addition, data mining is proposed for assisting physicians in knowledge extraction and decision making. The intelligent analysis of clinical data mainly through association analysis and clustering techniques is anticipated to provide the physicians with insight into an individual’s glycemic status and enables prediction of the long-term course of the disease and diabetic complications.

Keywords

diabetes glucose prediction machine learning data mining 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • E. I. Georga
    • 1
  • V. C. Protopappas
    • 1
  • C. Bellos
    • 1
  • D. Makriyiannis
    • 2
  • D. I. Fotiadis
    • 1
  1. 1.Department of Materials Science and EngineeringUniversity of IoanninaIoanninaGreece
  2. 2.Department of EndocrinologyHatzikosta General HospitalIoanninaGreece

Personalised recommendations