Glucose Prediction and Hypoglycemia Alarms Based on Adaptive Model

Abstract:

The paper proposes a glucose prediction model and hypoglycemia alarms technology based on CGMS. Method: By using kalman filter to smooth the glucose data from the CGMS, reducing noise interference; Then according to the non-stationary characteristics of glucose concentration signal ,Using adaptive autoregressive model (AR) glucose prediction model is established; Finally, the prediction model is applied to hypoglycemia alarms. Results: The prediction model can dynamically capture the changes of the glucose and predict glucose of 30 min ahead, RMSE、SSGPE were 5.069,5.276; And hypoglycemia can be timely detected.

• Periodical: Advanced Materials Research (Volumes 971-973)
• Pages: 275-279
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.971-973.275
• Online since: June 2014
• Authors: Yan Nian Wang*, Yan Rui Shen, Yong Qiang Yong, Quan Zhong Li, Chang Qing Sun
• Keywords: Adaptive System, Glucose Prediction, Hypoglycemia Alarm, Non-Stationary Characteristics, Smoothing Process

References

[1] Z. Trajanoski. Simulation studies on neural predictive control of glucose using the subcutaneous route[J], Comput. Methods Programs Biomed., 1998, 56: 133–139.

DOI: 10.1016/s0169-2607(98)00020-0

[2] C. Zecchin, A. Facchinetti, G. Sparacino, etal. A New Neural Network Approach for Short-Term Glucose Prediction Using Continuous Glucose Monitoring Time-Series and Meal Information[J]. 33rd Annual International Conference of the IEEE EMBS, Boston, Massachusetts USA, August 30 - September 3, 2011, 5653-5656.

DOI: 10.1109/iembs.2011.6091368

[3] Z Wang, L Lai, D Xiong, X Wu. Study on predicting method for acute hypotensive episodes based on wavelet transform and Support Vector Machine. BMEI[C], (2010).

DOI: 10.1109/bmei.2010.5639747

[4] G. Sparacino, F. Zanderigo, S. Corazza, A. Maran, A. Facchinetti, and C. Cobelli. Glucose concentration can be predicted ahead in time from continuous glucose monitoring sensor time-series, IEEE Trans. Biomed. Eng, 2007, 54(5): 931–937.

DOI: 10.1109/tbme.2006.889774

[5] Bayrak ES, Kamuran Turksoy, B. S, etal. Hypoglycemia Early Alarm Systems Based on Recursive Autoregressive Partial Least Squares Models[J]. Journal of Diabetes Science and Technology, 2013, 7(1): 206 – 214.

DOI: 10.1177/193229681300700126

[6] Khovanova NA, Khovanov IA, Sbano L, Griffiths F, Holt TA. Characterization of linear predictability and non-stationarity of subcutaneous glucose profiles[J]. Computer Methods and Programs in Biomedicine, 2013: 260 – 267.

DOI: 10.1016/j.cmpb.2012.11.009

[7] Yumin Pan, Liyong Zhao, Quanzhu Zhang etal. The stable prediction method based on high frequency random time sequence component [J]. Computer Engineering and Design, 2013, 34(6).

[8] Bequette, B. W. Continuous Glucose Monitoring: Real-Time Algorithms for Calibration, Filtering, and Alarms. [J]. Diabetes Sci. Technol. 2010, 4 (2), 404−18.

DOI: 10.1177/193229681000400222

[9] Facchinetti, A.; Sparacino, G.; Cobelli, C. An Online SelfTunable Method to Denoise CGM Sensor Data. IEEE Trans. Bio-Med. Eng. 2010, 57 (3), 634−641.

DOI: 10.1109/tbme.2009.2033264

[10] Giurcăneanu C D, Razavi S A. AR order selection in the case when the model parameters are estimated by forgetting factor least-squares algorithms[J]. Signal Processing, 2010, 90(2): 451-466.

DOI: 10.1016/j.sigpro.2009.07.011

[11] Eren-Oruklu M, Cinar A, Rollins D K, etal. Adaptive system identification for estimating future glucoseconcentrations and hypoglycemia alarms[J]. Automatica, 2012, 48(8): 1892-1897.

DOI: 10.1016/j.automatica.2012.05.076

[12] Turksoy K, Bayrak E S, Quinn L, etal. Hypoglycemia Early Alarm Systems Based on Multivariable Models[J]. Industrial & Engineering Chemistry Research, (2013).

DOI: 10.1021/ie3034015