Advertisement

Medical & Biological Engineering & Computing

, Volume 57, Issue 1, pp 27–46 | Cite as

Short-term prediction of glucose in type 1 diabetes using kernel adaptive filters

  • Eleni I. Georga
  • José C. Príncipe
  • Dimitrios I. FotiadisEmail author
Original Article
  • 129 Downloads

Abstract

This study aims at presenting a nonlinear, recursive, multivariate prediction model of the subcutaneous glucose concentration in type 1 diabetes. Nonlinear regression is performed in a reproducing kernel Hilbert space, by either the fixed budget quantized kernel least mean square (QKLMS-FB) or the approximate linear dependency kernel recursive least-squares (KRLS-ALD) algorithm, such that a sparse model structure is accomplished. A multivariate feature set (i.e., subcutaneous glucose, food carbohydrates, insulin regime and physical activity) is used and its influence on short-term glucose prediction is investigated. The method is evaluated using data from 15 patients with type 1 diabetes in free-living conditions. In the case when all the input variables are considered: (i) the average root mean squared error (RMSE) of QKLMS-FB increases from 13.1 mg dL−1 (mean absolute percentage error (MAPE) 6.6%) for a 15-min prediction horizon (PH) to 37.7 mg dL−1 (MAPE 20.8%) for a 60-min PH and (ii) the RMSE of KRLS-ALD, being predictably lower, increases from 10.5 mg dL−1 (MAPE 5.2%) for a 15-min PH to 31.8 mg dL−1 (MAPE 18.0%) for a 60-min PH. Multivariate data improve systematically both the regularity and the time lag of the predictions, reducing the errors in critical glucose value regions for a PH ≥ 30 min.

Graphical abstract

Keywords

Glucose concentration prediction Kernel methods Nonlinear regression Online learning Type 1 diabetes 

Notes

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent

Informed consent was obtained from all individual participants included in the study.

Supplementary material

11517_2018_1859_MOESM1_ESM.docx (17 kb)
ESM 1 (DOCX 16 kb)

References

  1. 1.
    Cescon M, Johansson R, Renard E (2015) Subspace-based linear multi-step predictors in type 1 diabetes mellitus. Biomed Signal Process Control 22:99–110.  https://doi.org/10.1016/j.bspc.2014.09.012 CrossRefGoogle Scholar
  2. 2.
    Chen BD, Zhao SL, Zhu PP, Principe JC (2012) Quantized kernel least mean square algorithm. Ieee T Neur Net Lear 23:22–32.  https://doi.org/10.1109/Tnnls.2011.2178446 CrossRefGoogle Scholar
  3. 3.
    Cryer PE (2009) Exercise-related hypoglycemia-associated autonomic failure in diabetes. Diabetes 58:1951–1952.  https://doi.org/10.2337/db09-0834 CrossRefPubMedPubMedCentralGoogle Scholar
  4. 4.
    Daskalaki E, Prountzou A, Diem P, Mougiakakou SG (2012) Real-time adaptive models for the personalized prediction of glycemic profile in type 1 diabetes patients. Diabetes Technol Ther 14:168–174.  https://doi.org/10.1089/dia.2011.0093 CrossRefPubMedGoogle Scholar
  5. 5.
    Engel Y, Mannor S, Meir R (2004) The kernel recursive least-squares algorithm. Ieee T Signal Proces 52:2275–2285.  https://doi.org/10.1109/Tsp.2004.830985 CrossRefGoogle Scholar
  6. 6.
    Eren-Oruklu M, Cinar A, Quinn L, Smith D (2009) Estimation of future glucose concentrations with subject-specific recursive linear models. Diabetes Technol Ther 11:243–253.  https://doi.org/10.1089/dia.2008.0065 CrossRefPubMedPubMedCentralGoogle Scholar
  7. 7.
    Eren-Oruklu M, Cinar A, Rollins DK, Quinn L (2012) Adaptive system identification for estimating future glucose concentrations and hypoglycemia alarms. Automatica 48:1892–1897.  https://doi.org/10.1016/j.automatica.2012.05.076
  8. 8.
    Facchinetti A, Sparacino G, Trifoglio E, Cobelli C (2011) A new index to optimally design and compare continuous glucose monitoring glucose prediction algorithms. Diabetes Technol Ther 13:111–119.  https://doi.org/10.1089/dia.2010.0151 CrossRefPubMedGoogle Scholar
  9. 9.
    Finan DA, Doyle FJ 3rd, Palerm CC, Bevier WC, Zisser HC, Jovanovic L, Seborg DE (2009) Experimental evaluation of a recursive model identification technique for type 1 diabetes. J Diabetes Sci Technol 3:1192–1202CrossRefGoogle Scholar
  10. 10.
    Finan DA, Palerm CC, Doyle FJ, Seborg DE, Zisser H, Bevier WC, Jovanovič L (2009) Effect of input excitation on the quality of empirical dynamic models for type 1 diabetes. AICHE J 55:1135–1146.  https://doi.org/10.1002/aic.11699 CrossRefGoogle Scholar
  11. 11.
    Gani A, Gribok AV, Lu Y, Ward WK, Vigersky RA, Reifman J (2010) Universal glucose models for predicting subcutaneous glucose concentration in humans. IEEE Trans Inf Technol Biomed 14:157–165.  https://doi.org/10.1109/TITB.2009.2034141
  12. 12.
    Gani A, Gribok AV, Rajaraman S, Ward WK, Reifman J (2009) Predicting subcutaneous glucose concentration in humans: data-driven glucose modeling. IEEE Trans Biomed Eng 56:246–254.  https://doi.org/10.1109/TBME.2008.2005937 CrossRefPubMedGoogle Scholar
  13. 13.
    Georga E, Protopappas V, Guillen A, Fico G, Ardigo D, Arredondo MT, Exarchos TP, Polyzos D, Fotiadis DI (2009) Data mining for blood glucose prediction and knowledge discovery in diabetic patients: the METABO diabetes modeling and management system. Conf Proc IEEE Eng Med Biol Soc 2009:5633–5636.  https://doi.org/10.1109/IEMBS.2009.5333635
  14. 14.
    Georga EI, Principe JC, Polyzos D, Fotiadis DI (2016) Non-linear dynamic modeling of glucose in type 1 diabetes with kernel adaptive filters. Conf Proc IEEE Eng Med Biol Soc, 16–20 Aug 2016. pp 5897–5900. doi: https://doi.org/10.1109/EMBC.2016.7592070
  15. 15.
    Georga EI, Príncipe JC, Rizos EC, Fotiadis DI (2017) Kernel-based adaptive learning improves accuracy of glucose predictive modelling in type 1 diabetes: a proof-of-concept study. In: 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 11–15 July 2017. pp 2765–2768. doi: https://doi.org/10.1109/EMBC.2017.8037430
  16. 16.
    Georga EI, Protopappas VC, Ardigò D, Marina M, Zavaroni I, Polyzos D, Fotiadis DI (2013) Multivariate prediction of subcutaneous glucose concentration in type 1 diabetes patients based on support vector regression. IEEE J Biomed Health Inform 17:71–81.  https://doi.org/10.1109/TITB.2012.2219876
  17. 17.
    Georga EI, Protopappas VC, Ardigo D, Polyzos D, Fotiadis DI (2013) A glucose model based on support vector regression for the prediction of hypoglycemic events under free-living conditions. Diabetes Technol Ther 15:634–643.  https://doi.org/10.1089/dia.2012.0285 CrossRefPubMedGoogle Scholar
  18. 18.
    Georga EI, Protopappas VC, Polyzos D, Fotiadis DI (2015) Evaluation of short-term predictors of glucose concentration in type 1 diabetes combining feature ranking with regression models. Med Biol Eng Comput 53:1305–1318.  https://doi.org/10.1007/s11517-015-1263-1 CrossRefPubMedGoogle Scholar
  19. 19.
    Glucose concentrations of less than 3.0 mmol/L (54 mg/dL) should be reported in clinical Trials: a joint position statement of the American Diabetes Association and the European Association for the Study of Diabetes (2017). Diabetes care 40:155–157. doi: https://doi.org/10.2337/dc16-2215
  20. 20.
    Kovatchev B, Clarke W (2008) Peculiarities of the continuous glucose monitoring data stream and their impact on developing closed-loop control technology. J Diabetes Sci Technol 2:158–163CrossRefGoogle Scholar
  21. 21.
    Kovatchev BP, Gonder-Frederick LA, Cox DJ, Clarke WL (2004) Evaluating the accuracy of continuous glucose-monitoring sensors: continuous glucose-error grid analysis illustrated by TheraSense freestyle navigator data. Diabetes Care 27:1922–1928CrossRefGoogle Scholar
  22. 22.
    Lehmann ED, Deutsch T (1992) A physiological model of glucose-insulin interaction in type 1 diabetes mellitus. J Biomed Eng 14:235–242CrossRefGoogle Scholar
  23. 23.
    Li K, Principe JC (2016) The kernel adaptive autoregressive-moving-average algorithm. IEEE Trans Neural Netw Learn Syst 27:334–346.  https://doi.org/10.1109/TNNLS.2015.2418323 CrossRefPubMedGoogle Scholar
  24. 24.
    Li K, Príncipe JC (2017) Transfer learning in adaptive filters: the nearest instance centroid-estimation kernel least-mean-square algorithm. Ieee T Signal Proces 65:6520–6535.  https://doi.org/10.1109/TSP.2017.2752695 CrossRefGoogle Scholar
  25. 25.
    Liu W, Park I, Principe JC (2009) An information theoretic approach of designing sparse kernel adaptive filters. IEEE Trans Neural Netw 20:1950–1961.  https://doi.org/10.1109/TNN.2009.2033676
  26. 26.
    Liu W, Príncipe JC, Haykin S (2010) Background and preview. In: Kernel Adaptive Filtering. John Wiley & Sons, Inc., pp 1–26. doi: https://doi.org/10.1002/9780470608593.ch1
  27. 27.
    Liu W, Príncipe JC, Haykin S (2010) Extended kernel recursive least-squares algorithm. In: Kernel Adaptive Filtering. John Wiley & Sons, Inc., pp 124–151. doi: https://doi.org/10.1002/9780470608593.ch5
  28. 28.
    Liu W, Príncipe JC, Haykin S (2010) Kernel recursive least-squares algorithm. In: Kernel Adaptive Filtering. John Wiley & Sons, Inc., pp 94–123. doi: https://doi.org/10.1002/9780470608593.ch4
  29. 29.
    Naumova V, Pereverzyev SV, Sivananthan S (2012) A meta-learning approach to the regularized learning-case study: blood glucose prediction. Neural Networks 33:181–193.  https://doi.org/10.1016/j.neunet.2012.05.004
  30. 30.
    Oviedo S, Vehi J, Calm R, Armengol J (2017) A review of personalized blood glucose prediction strategies for T1DM patients. 33. doi: https://doi.org/10.1002/cnm.2833
  31. 31.
    Pappada SM, Cameron BD, Rosman PM, Bourey RE, Papadimos TJ, Olorunto W, Borst MJ (2011) Neural network-based real-time prediction of glucose in patients with insulin-dependent diabetes. Diabetes Technol Ther 13:135–141.  https://doi.org/10.1089/dia.2010.0104 CrossRefPubMedGoogle Scholar
  32. 32.
    Perez-Gandia C, Facchinetti A, Sparacino G, Cobelli C, Gomez EJ, Rigla M, de Leiva A, Hernando ME (2010) Artificial neural network algorithm for online glucose prediction from continuous glucose monitoring. Diabetes Technol Ther 12:81–88.  https://doi.org/10.1089/dia.2009.0076 CrossRefPubMedGoogle Scholar
  33. 33.
    Pokharel R, Seth S, Principe JC Mixture kernel least mean square. In: The 2013 International Joint Conference on Neural Networks (IJCNN), 4–9 Aug. 2013 2013. pp 1–7. doi: https://doi.org/10.1109/IJCNN.2013.6706867
  34. 34.
    Reifman J, Rajaraman S, Gribok A, Ward WK (2007) Predictive monitoring for improved management of glucose levels. J Diabetes Sci Technol 1:478–486CrossRefGoogle Scholar
  35. 35.
    Shrayyef MZ, Gerich JE (2010) Normal glucose homeostasis. In: Poretsky L (ed) Principles of diabetes mellitus. Springer US, Boston, pp 19–35.  https://doi.org/10.1007/978-0-387-09841-8_2 CrossRefGoogle Scholar
  36. 36.
    Sparacino G, Zanderigo F, Corazza S, Maran A, Facchinetti A, Cobelli C (2007) Glucose concentration can be predicted ahead in time from continuous glucose monitoring sensior time-series. IEEE Trans Biomed Eng 54:931–937.  https://doi.org/10.1109/Tbme.2006.889774 CrossRefPubMedGoogle Scholar
  37. 37.
    Stahl F, Johansson R (2009) Diabetes mellitus modeling and short-term prediction based on blood glucose measurements. Math Biosci 217:101–117.  https://doi.org/10.1016/j.mbs.2008.10.008 CrossRefPubMedGoogle Scholar
  38. 38.
    Tarin C, Teufel E, Pico J, Bondia J, Pfleiderer HJ (2005) Comprehensive pharmacokinetic model of insulin glargine and other insulin formulations. IEEE Trans Biomed Eng 52:1994–2005.  https://doi.org/10.1109/TBME.2005.857681 CrossRefPubMedGoogle Scholar
  39. 39.
    Turksoy K, Bayrak ES, Quinn L, Littlejohn E, Rollins D, Cinar A (2013) Hypoglycemia early alarm systems based on multivariable models. Ind Eng Chem Res 52:12329–12336.  https://doi.org/10.1021/ie3034015 CrossRefGoogle Scholar
  40. 40.
    Turksoy K, Quinn L, Littlejohn E, Cinar A (2014) Multivariable adaptive identification and control for artificial pancreas systems. IEEE Trans Biomed Eng 61:883–891.  https://doi.org/10.1109/TBME.2013.2291777 CrossRefPubMedGoogle Scholar
  41. 41.
    Vaerenbergh SV, Santamaría I A comparative study of kernel adaptive filtering algorithms. In: 2013 IEEE Digital Signal Processing and Signal Processing Education Meeting (DSP/SPE), 11–14 Aug. 2013 2013. pp 181–186. doi: https://doi.org/10.1109/DSP-SPE.2013.6642587
  42. 42.
    Wang Q, Molenaar P, Harsh S, Freeman K, Xie J, Gold C, Rovine M, Ulbrecht J (2014) Personalized state-space modeling of glucose dynamics for type 1 diabetes using continuously monitored glucose, insulin dose, and meal intake: an extended Kalman filter approach. J Diabetes Sci Technol 8:331–345.  https://doi.org/10.1177/1932296814524080 CrossRefPubMedPubMedCentralGoogle Scholar
  43. 43.
    Wang Y, Wu X, Mo X (2013) A novel adaptive-weighted-average framework for blood glucose prediction. Diabetes Technol Ther 15:792–801.  https://doi.org/10.1089/dia.2013.0104 CrossRefPubMedPubMedCentralGoogle Scholar
  44. 44.
    Yu X, Turksoy K, Rashid M, Feng J, Hobbs N, Hajizadeh I, Samadi S, Sevil M, Lazaro C, Maloney Z, Littlejohn E, Quinn L, Cinar A (2018) Model-fusion-based online glucose concentration predictions in people with type 1 diabetes. Control Eng Pract 71:129–141.  https://doi.org/10.1016/j.conengprac.2017.10.013 CrossRefPubMedGoogle Scholar
  45. 45.
    Zarkogianni K, Mitsis K, Litsa E, Arredondo MT, Ficomicron G, Fioravanti A, Nikita KS (2015) Comparative assessment of glucose prediction models for patients with type 1 diabetes mellitus applying sensors for glucose and physical activity monitoring. Med Biol Eng Comput 53:1333–1343.  https://doi.org/10.1007/s11517-015-1320-9 CrossRefPubMedGoogle Scholar
  46. 46.
    Zecchin C, Facchinetti A, Sparacino G, Cobelli C (2014) Jump neural network for online short-time prediction of blood glucose from continuous monitoring sensors and meal information. Comput Methods Prog Biomed 113:144–152.  https://doi.org/10.1016/j.cmpb.2013.09.016 CrossRefGoogle Scholar
  47. 47.
    Zecchin C, Facchinetti A, Sparacino G, Cobelli C (2016) How much is short-term glucose prediction in type 1 diabetes improved by adding insulin delivery and meal content information to CGM data? A proof-of-concept study. J Diabetes Sci Technol 10:1149–1160.  https://doi.org/10.1177/1932296816654161 CrossRefPubMedPubMedCentralGoogle Scholar
  48. 48.
    Zecchin C, Facchinetti A, Sparacino G, De Nicolao G, Cobelli C (2012) Neural network incorporating meal information improves accuracy of short-time prediction of glucose concentration. IEEE Trans Biomed Eng 59:1550–1560.  https://doi.org/10.1109/TBME.2012.2188893 CrossRefPubMedGoogle Scholar
  49. 49.
    Zhao C, Dassau E, Jovanovic L, Zisser HC, Doyle FJ 3rd, Seborg DE (2012) Predicting subcutaneous glucose concentration using a latent-variable-based statistical method for type 1 diabetes mellitus. J Diabetes Sci Technol 6:617–633CrossRefGoogle Scholar
  50. 50.
    Zhao C, Sun Y, Zhao L (2013) Interindividual glucose dynamics in different frequency bands for online prediction of subcutaneous glucose concentration in type 1 diabetic subjects. AICHE J 59:4228–4240.  https://doi.org/10.1002/aic.14176 CrossRefGoogle Scholar
  51. 51.
    Zhao C, Yu C (2015) Rapid model identification for online subcutaneous glucose concentration prediction for new subjects with type I diabetes. IEEE Trans Biomed Eng 62:1333–1344.  https://doi.org/10.1109/TBME.2014.2387293 CrossRefPubMedGoogle Scholar
  52. 52.
    Zhao SL, Chen BD, Zhu PP, Principe JC (2013) Fixed budget quantized kernel least-mean-square algorithm. Signal Process 93:2759–2770.  https://doi.org/10.1016/j.sigpro.2013.02.012 CrossRefGoogle Scholar

Copyright information

© International Federation for Medical and Biological Engineering 2018

Authors and Affiliations

  1. 1.Unit of Medical Technology and Intelligent Information Systems, Department of Materials Science and EngineeringUniversity of IoanninaIoanninaGreece
  2. 2.Computational NeuroEngineering LaboratoryUniversity of FloridaGainesvilleUSA
  3. 3.Department of Biomedical ResearchInstitute of Molecular Biology and Biotechnology, Foundation for Research and Technology - Hellas (FORTH)IoanninaGreece

Personalised recommendations