{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,13]],"date-time":"2026-10-13T16:33:28Z","timestamp":1791909208732,"version":"4.3.4"},"reference-count":55,"publisher":"MDPI AG","issue":"5","license":[{"start":{"date-parts":[[2021,2,27]],"date-time":"2021-02-27T00:00:00Z","timestamp":1614384000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100003407","name":"Ministero dell\u2019Istruzione, dell\u2019Universit\u00e0 e della Ricerca","doi-asserted-by":"publisher","award":["RBSI14JYM2"],"award-info":[{"award-number":["RBSI14JYM2"]}],"id":[{"id":"10.13039\/501100003407","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"<jats:p>In type 1 diabetes management, the availability of algorithms capable of accurately forecasting future blood glucose (BG) concentrations and hypoglycemic episodes could enable proactive therapeutic actions, e.g., the consumption of carbohydrates to mitigate, or even avoid, an impending critical event. The only input of this kind of algorithm is often continuous glucose monitoring (CGM) sensor data, because other signals (such as injected insulin, ingested carbs, and physical activity) are frequently unavailable. Several predictive algorithms fed by CGM data only have been proposed in the literature, but they were assessed using datasets originated by different experimental protocols, making a comparison of their relative merits difficult. The aim of the present work was to perform a head-to-head comparison of thirty different linear and nonlinear predictive algorithms using the same dataset, given by 124 CGM traces collected over 10 days with the newest Dexcom G6 sensor available on the market and considering a 30-min prediction horizon. We considered the state-of-the art methods, investigating, in particular, linear black-box methods (autoregressive; autoregressive moving-average; and autoregressive integrated moving-average, ARIMA) and nonlinear machine-learning methods (support vector regression, SVR; regression random forest; feed-forward neural network, fNN; and long short-term memory neural network). For each method, the prediction accuracy and hypoglycemia detection capabilities were assessed using either population or individualized model parameters. As far as prediction accuracy is concerned, the results show that the best linear algorithm (individualized ARIMA) provides accuracy comparable to that of the best nonlinear algorithm (individualized fNN), with root mean square errors of 22.15 and 21.52 mg\/dL, respectively. As far as hypoglycemia detection is concerned, the best linear algorithm (individualized ARIMA) provided precision = 64%, recall = 82%, and one false alarm\/day, comparable to the best nonlinear technique (population SVR): precision = 63%, recall = 69%, and 0.5 false alarms\/day. In general, the head-to-head comparison of the thirty algorithms fed by CGM data only made using a wide dataset shows that individualized linear models are more effective than population ones, while no significant advantages seem to emerge when employing nonlinear methodologies.<\/jats:p>","DOI":"10.3390\/s21051647","type":"journal-article","created":{"date-parts":[[2021,2,27]],"date-time":"2021-02-27T03:44:31Z","timestamp":1614397471000},"page":"1647","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":68,"title":["Forecasting of Glucose Levels and Hypoglycemic Events: Head-to-Head Comparison of Linear and Nonlinear Data-Driven Algorithms Based on Continuous Glucose Monitoring Data Only"],"prefix":"10.3390","volume":"21","author":[{"given":"Francesco","family":"Prendin","sequence":"first","affiliation":[{"name":"Department of Information Engineering, University of Padova, 35131 Padova, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8214-2752","authenticated-orcid":false,"given":"Simone","family":"Del Favero","sequence":"additional","affiliation":[{"name":"Department of Information Engineering, University of Padova, 35131 Padova, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Martina","family":"Vettoretti","sequence":"additional","affiliation":[{"name":"Department of Information Engineering, University of Padova, 35131 Padova, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3248-1393","authenticated-orcid":false,"given":"Giovanni","family":"Sparacino","sequence":"additional","affiliation":[{"name":"Department of Information Engineering, University of Padova, 35131 Padova, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8041-2280","authenticated-orcid":false,"given":"Andrea","family":"Facchinetti","sequence":"additional","affiliation":[{"name":"Department of Information Engineering, University of Padova, 35131 Padova, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2021,2,27]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"37","DOI":"10.1016\/j.ecl.2019.10.006","article-title":"Diabetes Technology Use in Adults with Type 1 and Type 2 Diabetes","volume":"49","author":"Kravarusic","year":"2020","journal-title":"Endocrinol. Metab. Clin."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.ecl.2019.10.009","article-title":"Evolution of Diabetes Technology","volume":"49","author":"Dovc","year":"2020","journal-title":"Endocrinol. Metab. Clin."},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"175","DOI":"10.1016\/j.bios.2018.03.039","article-title":"In-vitro model for assessing glucose diffusion through skin","volume":"110","author":"Ullah","year":"2018","journal-title":"Biosens. Bioelectron."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"178","DOI":"10.1016\/j.diabres.2017.08.005","article-title":"Continuous glucose monitoring: A review of the technology and clinical use","volume":"133","author":"Klonoff","year":"2017","journal-title":"Diabetes Res. Clin. Pract."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"800","DOI":"10.3390\/s19040800","article-title":"The progress of glucose monitoring\u2014A review of invasive to minimally and non-invasive techniques, devices and sensors","volume":"19","author":"Mobashsher","year":"2019","journal-title":"Sensors"},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"6925","DOI":"10.3390\/s20236925","article-title":"Non-Invasive Blood Glucose Monitoring Technology: A Review","volume":"20","author":"Tang","year":"2020","journal-title":"Sensors"},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"789","DOI":"10.1177\/193229681300700324","article-title":"\u201cTurn it off!\u201d: Diabetes device alarm fatigue considerations for the present and the future","volume":"7","author":"Shivers","year":"2013","journal-title":"J. Diabetes Sci. Technol."},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"McGarraugh, G. (2010). Alarm characterization for continuous glucose monitors used as adjuncts to self-monitoring of blood glucose. JDST.","DOI":"10.1177\/193229681000400106"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"383","DOI":"10.4093\/dmj.2019.0121","article-title":"Continuous glucose monitoring sensors for diabetes management: A review of technologies and applications","volume":"43","author":"Cappon","year":"2019","journal-title":"Diabetes Metab. J."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"644","DOI":"10.1089\/dia.2019.0139","article-title":"A Real-Time Continuous Glucose Monitoring\u2013Based Algorithm to Trigger Hypotreatments to Prevent\/Mitigate Hypoglycemic Events","volume":"21","author":"Camerlingo","year":"2019","journal-title":"Diabetes Technol. Ther."},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"2633","DOI":"10.1109\/JBHI.2018.2887067","article-title":"A dual mode adaptive basal-bolus advisor based on reinforcement learning","volume":"23","author":"Sun","year":"2018","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"3","DOI":"10.1089\/dia.2005.7.3","article-title":"Hypoglycemia prediction and detection using optimal estimation","volume":"7","author":"Palerm","year":"2005","journal-title":"Diabetes Technol. Ther."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"624","DOI":"10.1177\/193229680700100505","article-title":"Hypoglycemia detection and prediction using continuous glucose monitoring-a study on hypoglycemic clamp data","volume":"1","author":"Palerm","year":"2007","journal-title":"J. Diabetes Sci. Technol."},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"1251","DOI":"10.1109\/JBHI.2018.2840690","article-title":"An ARIMA model with adaptive orders for predicting blood glucose concentrations and hypoglycemia","volume":"23","author":"Yang","year":"2018","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"ref_15","unstructured":"Zhu, T., Li, K., Herrero, P., Chen, J., and Georgiou, P. (2021, January 22). A Deep Learning Algorithm for Personalized Blood Glucose Prediction. KHD@ IJCAI. Available online: http:\/\/ceur-ws.org\/Vol-2148\/paper12.pdf."},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Sun, Q., Jankovic, M.V., Bally, L., and Mougiakakou, S.G. (2018, January 20\u201321). Predicting blood glucose with an LSTM and Bi-LSTM based deep neural network. Proceedings of the 2018 14th Symposium on Neural Networks and Applications (NEUREL), Belgrade, Serbia.","DOI":"10.1109\/NEUREL.2018.8586990"},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"144","DOI":"10.1016\/j.cmpb.2013.09.016","article-title":"Jump neural network for online short-time prediction of blood glucose from continuous monitoring sensors and meal information","volume":"113","author":"Zecchin","year":"2014","journal-title":"Comput. Methods Programs Biomed."},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"142","DOI":"10.1007\/s10916-017-0788-2","article-title":"Data based prediction of blood glucose concentrations using evolutionary methods","volume":"41","author":"Hidalgo","year":"2017","journal-title":"J. Med. Syst."},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"e2833","DOI":"10.1002\/cnm.2833","article-title":"A review of personalized blood glucose prediction strategies for T1DM patients","volume":"33","author":"Oviedo","year":"2017","journal-title":"Int. J. Numer. Methods Biomed. Eng."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"1149","DOI":"10.1177\/1932296816654161","article-title":"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","volume":"10","author":"Zecchin","year":"2016","journal-title":"J. Diabetes Sci. Technol."},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"603","DOI":"10.1109\/JBHI.2019.2908488","article-title":"Convolutional recurrent neural networks for glucose prediction","volume":"24","author":"Li","year":"2019","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"31","DOI":"10.3390\/diagnostics9010031","article-title":"Current diabetes technology: Striving for the artificial pancreas","volume":"9","author":"Allen","year":"2019","journal-title":"Diagnostics"},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"2735","DOI":"10.1109\/TBME.2015.2470521","article-title":"A review of emerging technologies for the management of diabetes mellitus","volume":"62","author":"Zarkogianni","year":"2015","journal-title":"IEEE Trans. Biomed. Eng."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"157","DOI":"10.1109\/TITB.2009.2034141","article-title":"Universal glucose models for predicting subcutaneous glucose concentration in humans","volume":"14","author":"Gani","year":"2009","journal-title":"IEEE Trans. Inf. Technol. Biomed."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"1013","DOI":"10.2337\/dc09-2303","article-title":"Prevention of nocturnal hypoglycemia using predictive alarm algorithms and insulin pump suspension","volume":"33","author":"Buckingham","year":"2010","journal-title":"Diabetes Care"},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"1249","DOI":"10.2337\/dc09-1487","article-title":"Real-time hypoglycemia prediction suite using continuous glucose monitoring: A safety net for the artificial pancreas","volume":"33","author":"Dassau","year":"2010","journal-title":"Diabetes Care"},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"1333","DOI":"10.1007\/s11517-015-1320-9","article-title":"Comparative assessment of glucose prediction models for patients with type 1 diabetes mellitus applying sensors for glucose and physical activity monitoring","volume":"53","author":"Zarkogianni","year":"2015","journal-title":"Med. Biol. Eng. Comput."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41598-017-06478-4","article-title":"Chaotic time series prediction for glucose dynamics in type 1 diabetes mellitus using regime-switching models","volume":"7","author":"Frandes","year":"2017","journal-title":"Sci. Rep."},{"key":"ref_29","unstructured":"Joedicke, D., Garnica, O., Kronberger, G., Colmenar, J.M., Winkler, S., Velasco, J.M., Contador, S., and Hidalgo, J.I. (2020, August 29). Analysis of the performance of Genetic Programming on the Blood Glucose Level Prediction Challenge 2020. Available online: http:\/\/ceur-ws.org\/Vol-2675\/paper25.pdf."},{"key":"ref_30","unstructured":"McShinsky, R., and Marshall, B. (2020, August 29). Comparison of Forecasting Algorithms for Type 1 Diabetic Glucose Prediction on 30 and 60-Min Prediction Horizons. Available online: http:\/\/ceur-ws.org\/Vol-2675\/paper2.pdf."},{"key":"ref_31","unstructured":"Ljung, L. (1999). System Identification-Theory for the User, Prentice Hall PTR. [2nd ed.]."},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"657","DOI":"10.1016\/j.automatica.2014.01.001","article-title":"Kernel methods in system identification, machine learning and function estimation: A survey","volume":"50","author":"Pillonetto","year":"2014","journal-title":"Automatica"},{"key":"ref_33","doi-asserted-by":"crossref","first-page":"931","DOI":"10.1109\/TBME.2006.889774","article-title":"Glucose concentration can be predicted ahead in time from continuous glucose monitoring sensor time-series","volume":"54","author":"Sparacino","year":"2007","journal-title":"IEEE Trans. Biomed. Eng."},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"793","DOI":"10.2337\/dc12-0736","article-title":"Real-time improvement of continuous glucose monitoring accuracy: The smart sensor concept","volume":"36","author":"Facchinetti","year":"2013","journal-title":"Diabetes Care"},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"111","DOI":"10.1089\/dia.2010.0151","article-title":"A new index to optimally design and compare continuous glucose monitoring glucose prediction algorithms","volume":"13","author":"Facchinetti","year":"2011","journal-title":"Diabetes Technol. Ther."},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"8340","DOI":"10.3182\/20110828-6-IT-1002.01929","article-title":"A novel nonparametric approach for the identification of the glucose-insulin system in Type 1 diabetic patients","volume":"44","author":"Pillonetto","year":"2011","journal-title":"IFAC Proc. Vol."},{"key":"ref_37","doi-asserted-by":"crossref","first-page":"1142","DOI":"10.1177\/193229681200600519","article-title":"Inpatient studies of a Kalman-filter-based predictive pump shutoff algorithm","volume":"6","author":"Cameron","year":"2012","journal-title":"J. Diabetes Sci. Technol."},{"key":"ref_38","doi-asserted-by":"crossref","first-page":"650","DOI":"10.1109\/JBHI.2018.2823763","article-title":"Prediction of adverse glycemic events from continuous glucose monitoring signal","volume":"23","author":"Gadaleta","year":"2018","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"ref_39","doi-asserted-by":"crossref","first-page":"634","DOI":"10.1089\/dia.2012.0285","article-title":"A glucose model based on support vector regression for the prediction of hypoglycemic events under free-living conditions","volume":"15","author":"Georga","year":"2013","journal-title":"Diabetes Technol. Ther."},{"key":"ref_40","doi-asserted-by":"crossref","unstructured":"Bunescu, R., Struble, N., Marling, C., Shubrook, J., and Schwartz, F. (2013, January 4\u20137). Blood glucose level prediction using physiological models and support vector regression. Proceedings of the 2013 12th International Conference on Machine Learning and Applications, Miami, FL, USA.","DOI":"10.1109\/ICMLA.2013.30"},{"key":"ref_41","doi-asserted-by":"crossref","first-page":"4482","DOI":"10.3390\/s19204482","article-title":"Utility of Big Data in Predicting Short-Term Blood Glucose Levels in Type 1 Diabetes Mellitus Through Machine Learning Techniques","volume":"19","author":"Chatzigiannakis","year":"2019","journal-title":"Sensors"},{"key":"ref_42","doi-asserted-by":"crossref","first-page":"81","DOI":"10.1089\/dia.2009.0076","article-title":"Artificial neural network algorithm for online glucose prediction from continuous glucose monitoring","volume":"12","author":"Facchinetti","year":"2010","journal-title":"Diabetes Technol. Ther."},{"key":"ref_43","doi-asserted-by":"crossref","first-page":"69311","DOI":"10.1109\/ACCESS.2019.2919184","article-title":"A multi-patient data-driven approach to blood glucose prediction","volume":"7","author":"Aliberti","year":"2019","journal-title":"IEEE Access"},{"key":"ref_44","doi-asserted-by":"crossref","unstructured":"El Idriss, T., Idri, A., Abnane, I., and Bakkoury, Z. (2019, January 1\u20134). Predicting Blood Glucose using an LSTM Neural Network. Proceedings of the 2019 Federated Conference on Computer Science and Information Systems (FedCSIS), Leipzig, Germany.","DOI":"10.15439\/2019F159"},{"key":"ref_45","doi-asserted-by":"crossref","first-page":"309","DOI":"10.1080\/02664760903002667","article-title":"Support vector regression based residual control charts","volume":"37","author":"Gani","year":"2010","journal-title":"J. Appl. Stat."},{"key":"ref_46","doi-asserted-by":"crossref","first-page":"643","DOI":"10.1016\/S0925-2312(02)00632-X","article-title":"Determination of the spread parameter in the Gaussian kernel for classification and regression","volume":"55","author":"Wang","year":"2003","journal-title":"Neurocomputing"},{"key":"ref_47","doi-asserted-by":"crossref","first-page":"1947","DOI":"10.1021\/ci034160g","article-title":"Random forest: A classification and regression tool for compound classification and QSAR modeling","volume":"43","author":"Svetnik","year":"2003","journal-title":"J. Chem. Inf. Comput. Sci."},{"key":"ref_48","doi-asserted-by":"crossref","first-page":"349","DOI":"10.1081\/DDC-120018209","article-title":"Comparison of neural network and multiple linear regression as dissolution predictors","volume":"29","author":"Sathe","year":"2003","journal-title":"Drug Dev. Ind. Pharm."},{"key":"ref_49","doi-asserted-by":"crossref","unstructured":"Siami-Namini, S., Tavakoli, N., and Namin, A.S. (2018, January 17\u201320). A comparison of ARIMA and LSTM in forecasting time series. Proceedings of the 2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA), Orlando, FL, USA.","DOI":"10.1109\/ICMLA.2018.00227"},{"key":"ref_50","unstructured":"Hastie, T., Tibshirani, R., and Friedman, J. (2009). The Elements of Statistical Learning: Data Mining, Inference, and Prediction, Springer Science & Business Media."},{"key":"ref_51","doi-asserted-by":"crossref","unstructured":"Mirshekarian, S., Shen, H., Bunescu, R., and Marling, C. (2019, January 23\u201327). LSTMs and neural attention models for blood glucose prediction: Comparative experiments on real and synthetic data. Proceedings of the 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Berlin, Germany.","DOI":"10.1109\/EMBC.2019.8856940"},{"key":"ref_52","doi-asserted-by":"crossref","unstructured":"Saito, T., and Rehmsmeier, M. (2015). The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLoS ONE, 10.","DOI":"10.1371\/journal.pone.0118432"},{"key":"ref_53","doi-asserted-by":"crossref","first-page":"395","DOI":"10.1089\/dia.2018.0150","article-title":"Accuracy of a factory-calibrated, real-time continuous glucose monitoring system during 10 days of use in youth and adults with diabetes","volume":"20","author":"Wadwa","year":"2018","journal-title":"Diabetes Technol. Ther."},{"key":"ref_54","doi-asserted-by":"crossref","first-page":"689","DOI":"10.1177\/193229681300700314","article-title":"An early warning system for hypoglycemic\/hyperglycemic events based on fusion of adaptive prediction models","volume":"7","author":"Daskalaki","year":"2013","journal-title":"J. Diabetes Sci. Technol."},{"key":"ref_55","unstructured":"Frandes, M., Timar, B., and Lungeanu, D. (2021, January 22). A Risk Based Neural Network Approach for Predictive Modeling of Blood Glucose Dynamics. MIE, Available online: https:\/\/pubmed.ncbi.nlm.nih.gov\/27577449\/."}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/21\/5\/1647\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,11]],"date-time":"2025-10-11T05:29:45Z","timestamp":1760160585000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/21\/5\/1647"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,2,27]]},"references-count":55,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2021,3]]}},"alternative-id":["s21051647"],"URL":"https:\/\/doi.org\/10.3390\/s21051647","relation":{},"ISSN":["1424-8220"],"issn-type":[{"value":"1424-8220","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,2,27]]}}}