نوع مقاله : مقاله کامل پژوهشی

نویسندگان

1 کارشناسی ارشد، گروه مهندسی کامپیوتر، دانشکده‌ی مهندسی، دانشگاه گیلان، رشت، ایران

2 کارشناسی، گروه مهندسی کامپیوتر ، دانشکده‌ی مهندسی، دانشگاه گیلان، رشت، ایران

3 دانشیار، گروه مهندسی کامپیوتر، دانشکده‌ی مهندسی، دانشگاه گیلان، رشت، ایران

10.22041/ijbme.2020.134866.1617

چکیده

دیابت یک بیماری شایع در سراسر جهان است. این بیماری، سخت، غیر‌قابل علاج و در عین حال قابل کنترل بوده و از این رو کنترل و پیش‌گیری از عوارض آن امری مهم است. به همین دلیل استفاده از روش‌های هوشمند با خطای پایین برای پیش‌بینی میزان قند خون و از همه مهم‌تر جلوگیری از عوارض خطرناک آن یک مساله‌ی مهم در کنترل بهتر این بیماری است. با توجه به روش‌های مختلف ارائه‌ شده در این زمینه، در این مقاله نیز دو مدل با استفاده از ره‌یافت یادگیری عمیق ارائه شده که نتایج آن کارآمد و بهینه است. این دو مدل‌‌ پیشنهادی از ترکیب‌های متفاوتی از شبکه‌های عصبی حافظه‌ی طولانی کوتاه‌مدت و پیش‌خور تشکیل ‌شده و میزان قند خون آتی بیمار را با دقت و سرعت قابل توجهی پیش‌بینی می‌کنند. در این راستا از 81.200 داده‌ی میزان قند خون 203 بیمار به همراه 27 مشخصه‌ی موثر بر میزان قند خون استفاده شده است. هم‌چنین به منظور ارزیابی دقیق از روش اعتبارسنجی متقابل مناسب برای سری زمانی استفاده شده و نتایج حاصل از اجرای مدل‌ها نشان داده که مدل میانگین متحرک خودهمبسته‌ی یک‌پارچه با توجه به این حجم از داده و ضعف سخت‌افزاری سیستم پیاده‌سازی شده قادر به پیش‌بینی میزان قند خون نبوده در حالی که کارایی و سرعت عمل‌کرد مدل‌های مبتنی بر یادگیری عمیق قابل قبول است. هم‌چنین با توجه به نتایج به دست آمده مدل‌ پیشنهادی دوم برای افق‌های پیش‌بینی 5، 10 و 15 دقیقه به ترتیب 8/13%، 16% و 9/18% بهتر از مدل پیشنهادی اول عمل کرده و مدل قابل اعتمادتری برای پیش‌بینی میزان قند خون است. از این رو مدل‌ پیشنهادی دوم می‌تواند در سیستم‌های هوشمند هشداردهنده برای پیش‌گیری از وقوع هیپوگلیسمی که از عوارض خطرناک و شایع بیماری دیابت نوع یک است مورد استفاده قرار گیرد.

کلیدواژه‌ها

موضوعات

عنوان مقاله [English]

Blood Glucose Level Prediction for Type 1 Diabetes using Deep Learning

نویسندگان [English]

  • Seyedeh Sadaf Razavinezhad 1
  • Amir mohammad Fallah 2
  • Seyed Abolghasem Mirroshandel 3

1 M.Sc., Department of Computer Engineering, Faculty of Engineering, University of Guilan, Rasht, Iran

2 B.Sc., Department of Computer Engineering, Faculty of Engineering, University of Guilan, Rasht, Iran

3 Associate Professor, Department of Computer Engineering, Faculty of Engineering, University of Guilan, Rasht, Iran

چکیده [English]

Diabetes is a common disease all around the world. It is a difficult and incurable but controllable disease, so it is important to control and prevent its complications. Thus, low error and smart methods are used to predict blood glucose levels and prevent dangerous complications to control it effectively. In this regard, different methods were investigated. In this research, two models based on deep learning technique are used which produce efficient and optimal results. These models are composed of different combinations of long short-term memory and feed forward neural networks which predict the patient's future blood glucose levels with considerable accuracy and speed. To achieve more comprehensive model, 81,200 blood glucose data was evaluated through 203 patients. In addition, 27 effective features in patients' blood glucose levels are considered in this regard. Furthermore, cross-validation method which is suitable for time series was used for more accurate evaluation. The results showed that Autoregressive Integrated Moving Average model could not predict blood glucose levels considering this amount of data and system hardware limitations, while the models based on deep learning had good performance and good speed. Furthermore, the second proposed model for the prediction horizons of 5, 10, and 15 minutes outperformed the first one with 13.8%, 16%, and 18.9%, respectively. Therefore, the second proposed model can be more reliable for predicting blood glucose. So, it can be used in smart warning systems to prevent hypoglycemia, which is a dangerous and widespread problem of type 1 diabetes.

کلیدواژه‌ها [English]

  • blood glucose
  • Type 1 Diabetes
  • prediction
  • Artificial neural network
  • Deep Learning
  1. Ahmed, "History of diabetes mellitus," Saudi medical journal, vol. 23, no. 4, pp. 373-378, Apr 2002.
  2. Plis, R. Bunescu, C. Marling, J. Shubrook and F. Schwartz, "A machine learning approach to predicting blood glucose levels for diabetes management," in In Workshops at the Twenty-Eighth AAAI conference on artificial intelligence, 2014.
  3. E. Cryer, S. N. Davis and H. Shamoon, "Hypoglycemia in diabetes," Diabetes care, vol. 26, no. 6, pp. 1902-1912, Jun 2003.
  4. M. Dall, S. E. Mann, Y. Zhang, W. W. Quick, R. F. Seifert, J. Martin, E. A. Huang and S. Zhang, "Distinguishing the economic costs associated with type 1 and type 2 diabetes," Population health management, vol. 12, no. 2, pp. 103-110, 2009.
  5. B. Hirsch, R. Farkas-Hirsch and J. S. Skyler, "Intensive insulin therapy for treatment of type I diabetes," Diabetes Care, vol. 13, no. 12, pp. 1256-1283, Dec 1990.
  6. 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 Transactions on Biomedical Engineering, vol. 54, no. 5, pp. 931-937, Apr 2007.
  7. W. Hipel, A. l. McLeod and W. C. Lennox, "Advances in Box‐Jenkins modeling: 1. Model construction," Water Resources Research, vol. 13, no. 3, pp. 567-575, Jun 1977.
  8. L. Ho, M. Xie and T. N. Goh, "A comparative study of neural network and Box-Jenkins ARIMA modeling in time series prediction," Computers and Industrial Engineering, vol. 42, no. 2-4, pp. 371-375, Apr 2002.
  9. Bremer and D. A. Gough, "Is blood glucose predictable from previous values? A solicitation for data," Diabetes, vol. 48, no. 3, pp. 445-451, Mar 1999.
  10. Zanderigo, G. Sparacino, B. Kovatchev and C. Cobelli, "Glucose prediction algorithms from continuous monitoring data: assessment of accuracy via continuous glucose error-grid analysis," Journal of Diabetes Science and Technology, vol. 1, no. 5, pp. 645-651, Sep 2007.
  11. Eren-Oruklu, A. Cinar, L. Quinn and D. Smith, "Estimation of future glucose concentrations with subject-specific recursive linear models," Diabetes Technology and Therapeutics, vol. 11, no. 4, pp. 243-253, Apr 2009.
  12. Gani, A. V. Gribok, S. Rajaraman, W. K. Ward and J. Reifman, "Predicting subcutaneous glucose concentration in humans: data-driven glucose modeling," IEEE Transactions on Biomedical Engineering, vol. 56, no. 2, pp. 246-254, Feb 2008.
  13. Gani, A. V. Gribok, Y. Lu, W. K. Ward, R. A. Vigersky and J. Reifman, "Universal glucose models for predicting subcutaneous glucose concentration in humans," IEEE Transactions on Information Technology in Biomedicine, vol. 14, no. 1, pp. 157-165, Jan 2010.
  14. Hollnagel, The reliability of expert systems, First ed., Prentice-Hall, Inc., 1989, p. 243.
  15. Reifman, S. Rajaraman, A. Gribok and W. K. Ward, "Predictive monitoring for improved management of glucose levels," Diabetes Science and Technology, vol. 1, no. 4, pp. 478-486, Jul 2007.
  16. D. Cryer, Time series analysis, Boston: Duxbury Press, 1986, pp. 271-278.
  17. Kovatchev and W. Clarke, "Peculiarities of the continuous glucose monitoring data stream and their impact on developing closed-loop control technology," Diabetes and Technology, vol. 2, no. 1, pp. 158-163, Jan 2008.
  18. Ståhl and R. Johansson, "Diabetes mellitus modeling and short-term prediction based on blood glucose measurements," Mathematical biosciences, vol. 217, no. 2, pp. 101-117, Feb 2009.
  19. Hollnagel, Cognitive Reliability and Error Analysis Method (CREAM), First ed., Halden: Elsevier, 1998, pp. 1-21.
  20. Andrea and C. Cobelli, "Modeling Methodology for Physiology and Medicine," in Tracer Experiment Design for Metabolic Fluxes Estimation in Steady and Nonsteady State, Academic Press, 2001, pp. 153-178.
  21. G. Mougiakakou, A. Prountzou, D. Lliopoulou, K. S. Nikita, A. Vazeou and B. Christos S., "Neural network based glucose-insulin metabolism models for children with type 1 diabetes," in 2006 International Conference of the IEEE Engineering in Medicine and Biology Society, NY, 2006.
  22. J. Valletta, A. J. Chipperfield and C. D. Byrne, "Gaussian process modelling of blood glucose response to free-living physical activity data in people with type 1 diabetes," in 2009 Annual International Conference of the IEEE Engineer, MN, 2009.
  23. Pérez-Gandía, A. Facchinetti, G. Sparacino, C. Cobelli, E. J. Gomez, M. Rigla, A. d. Leiva and M. E. Hernado, "Neural Network Algorithm for Online Glucose Prediction from Continuous Glucose Monitoring," Diabetes technology & therapeutics, vol. 12, no. 1, pp. 81-88, Jan 2010.
  24. M. Pappada, B. D. Cameron and P. M. Rosman, "Development of a neural network for prediction of glucose concentration in type 1 diabetes patients," Diabetes Science and Technology, vol. 2, no. 5, pp. 792-801, Sep 2008.
  25. M. Pappada, B. D. Cameron, P. M. Rosman, R. E. Bourey, T. J. Papadimos, W. Olorunto and M. J. Borst, "Neural network-based real-time prediction of glucose in patients with insulin-dependent diabetes," Diabetes Technology and Therapeutics, vol. 13, no. 2, pp. 135-141, Feb 2011.
  26. Doike, K. Hayashi, S. Arata, K. N. Mohammad, A. Kobayashi and K. Niitsu, "A Blood Glucose Level Prediction System Using Machine Learning Based on Recurrent Neural Network for Hypoglycemia Prevention," in 2018 16th IEEE International New, QC, 2018.
  27. Li, J. Daniels, C. Liu, P. Herrero and P. Georgiou, "Convolutional recurrent neural networks for glucose prediction," IEEE Journal of Biomedical and Health Informatics, vol. 24, no. 2, pp. 603-613, Apr 2019.
  28. Sun, M. V. Jankovic, L. Bally and S. Mougiakakou G., "Predicting blood glucose with an LSTM and Bi-LSTM based deep neural network," in 2018 14th Symposium on Neural Networks and Applications (NEUREL), pp. 1-5, Nov 2018.
  29. Martinsson, S. Alexander, B. Eliasson and O. Morgen, "Blood glucose prediction with variance estimation using recurrent neural networks," Journal of Healthcare Informatics Research, vol. 4, no. 1, pp. 1-18, 2020.
  30. Munoz-organero, "Deep Physiological Model for Blood Glucose Prediction in T1DM Patients," Sensors, vol. 20, no. 14, p. 3896, Jan 2020.
  31. Sivananthan, V. Naumova, C. Dall Man, A. Facchinetti, E. Renard, C. Cobelli and S. V. Pereverzyev, "Assessment of Blood Glucose Predictors: The Prediction-Error Grid Analysis," DIABETES TECHNOLOGY & THERAPEUTICS, vol. 13, no. 8, pp. 787-796, Aug 2011.
  32. N. Mhaskar, S. V. Pereverzyev and M. D. van der Walt, "A deep learning approach to diabetic blood glucose prediction," Frontiers in Applied Mathematics and Statistics, vol. 3, p. 14, Jul 2017.
  33. Zecchin, A. Facchinetti, G. Sparacino and C. Cobelli, "Jump neural network for online short-time prediction of blood glucose from continuous monitoring sensors and meal information," Computer methods and programs in biomedicine, vol. 113, no. 1, pp. 144-152, Jan 2014.
  34. Kharazihai Isfahani, M. Zakeri, H. R. Marateb and E. Faghihimani, "A hybrid dynamic wavelet-based modeling method for blood glucose concentration prediction in type 1 diabetes," Medical Signals and Sensors, vol. 10, no. 3, pp. 174-184, Jul 2020.
  35. Xie and Q. Wang, "Benchmarking machine learning algorithms on blood glucose prediction for Type 1 Diabetes in comparison with classical time-series models," IEEE Transactions on Biomedical Engineering, 24 Feb 2020.
  36. Deng and Y. Dong, "Deep learning: methods and applications," in Foundations and Trends in Signal Processing, WA, 2014.
  37. C. Chen, G. Papandreou, F. Schroff and H. Adam, "Rethinking atrous convolution for semantic image segmentation," Jun 2017.
  38. M. Breuel, "The effects of hyperparameters on SGD training of neural networks," Aug 2015.