Evaluation of short-term predictors of glucose concentration in type 1 diabetes combining feature ranking with regression models

被引:59
作者
Georga, Eleni I. [1 ]
Protopappas, Vasilios C. [1 ]
Polyzos, Demosthenes [2 ]
Fotiadis, Dimitrios I. [1 ]
机构
[1] Univ Ioannina, Unit Med Technol & Intelligent Informat Syst, Dept Mat Sci & Engn, GR-45110 Ioannina, Greece
[2] Univ Patras, Dept Aeronaut & Mech Engn, Patras 26500, Greece
关键词
Type; 1; diabetes; Feature selection; Glucose level prediction; Machine learning; BLOOD-GLUCOSE; ARTIFICIAL PANCREAS; HYPOGLYCEMIC EVENTS; CIRCADIAN-RHYTHMS; NEURAL-NETWORK; IDENTIFICATION; CLASSIFICATION; REDUCTION; SELECTION; TUTORIAL;
D O I
10.1007/s11517-015-1263-1
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
Glucose concentration in type 1 diabetes is a function of biological and environmental factors which present high inter-patient variability. The objective of this study is to evaluate a number of features, which are extracted from medical and lifestyle self-monitoring data, with respect to their ability to predict the short-term subcutaneous (s.c.) glucose concentration of an individual. Random forests (RF) and RReliefF algorithms are first employed to rank the candidate feature set. Then, a forward selection procedure follows to build a glucose predictive model, where features are sequentially added to it in decreasing order of importance. Predictions are performed using support vector regression or Gaussian processes. The proposed method is validated on a dataset of 15 type diabetics in real-life conditions. The s.c. glucose profile along with time of the day and plasma insulin concentration are systematically highly ranked, while the effect of food intake and physical activity varies considerably among patients. Moreover, the average prediction error converges in less than d/2 iterations (d is the number of features). Our results suggest that RF and RReliefF can find the most informative features and can be successfully used to customize the input of glucose models.
引用
收藏
页码:1305 / 1318
页数:14
相关论文
共 48 条
[1]   A robust sliding mode controller with internal model for closed-loop artificial pancreas [J].
Abu-Rmileh, Amjad ;
Garcia-Gabin, Winston ;
Zambrano, Darine .
MEDICAL & BIOLOGICAL ENGINEERING & COMPUTING, 2010, 48 (12) :1191-1201
[2]   Selection bias in gene extraction on the basis of microarray gene-expression data [J].
Ambroise, C ;
McLachlan, GJ .
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 2002, 99 (10) :6562-6566
[3]   Standards of Medical Care in Diabetes-2014 [J].
不详 .
DIABETES CARE, 2014, 37 :S14-S80
[4]   Threshold-Based Insulin-Pump Interruption for Reduction of Hypoglycemia [J].
Bergenstal, Richard M. ;
Klonoff, David C. ;
Garg, Satish K. ;
Bode, Bruce W. ;
Meredith, Melissa ;
Slover, Robert H. ;
Ahmann, Andrew J. ;
Welsh, John B. ;
Lee, Scott W. ;
Kaufman, Francine R. .
NEW ENGLAND JOURNAL OF MEDICINE, 2013, 369 (03) :224-232
[5]   Random forests [J].
Breiman, L .
MACHINE LEARNING, 2001, 45 (01) :5-32
[6]   Outpatient Safety Assessment of an In-Home Predictive Low-Glucose Suspend System with Type 1 Diabetes Subjects at Elevated Risk of Nocturnal Hypoglycemia [J].
Buckingham, Bruce A. ;
Cameron, Fraser ;
Calhoun, Peter ;
Maahs, David M. ;
Wilson, Darrell M. ;
Chase, H. Peter ;
Bequette, B. Wayne ;
Lum, John ;
Sibayan, Judy ;
Beck, Roy W. ;
Kollman, Craig .
DIABETES TECHNOLOGY & THERAPEUTICS, 2013, 15 (08) :622-627
[7]   The Barrier of Hypoglycemia in Diabetes [J].
Cryer, Philip E. .
DIABETES, 2008, 57 (12) :3169-3176
[8]  
Daskalaki Elena, 2013, J Diabetes Sci Technol, V7, P689
[9]   The effect of physical exercise on the dynamics of glucose and insulin [J].
Derouich, M ;
Boutayeb, A .
JOURNAL OF BIOMECHANICS, 2002, 35 (07) :911-917
[10]   Effect of Size and Heterogeneity of Samples on Biomarker Discovery: Synthetic and Real Data Assessment [J].
Di Camillo, Barbara ;
Sanavia, Tiziana ;
Martini, Matteo ;
Jurman, Giuseppe ;
Sambo, Francesco ;
Barla, Annalisa ;
Squillario, Margherita ;
Furlanello, Cesare ;
Toffolo, Gianna ;
Cobelli, Claudio .
PLOS ONE, 2012, 7 (03)