Continuous time recurrent neural networks: Overview and benchmarking at forecasting blood glucose in the intensive care unit

Glu-Ensemble: An ensemble deep learning framework for blood glucose forecasting in type 2 diabetes patients

“…Fitzgerald et al [ 39 ] focused on the development and evaluation of continuous-time recurrent neural networks (CTRNNs) for forecasting blood glucose levels in intensive care unit (ICU) settings. The effectiveness of these models, which were designed to handle irregularly timed data inputs, was compared with that of traditional autoregressive models and gradient boosted trees (GBTs).…”

“…This, in turn, facilitates better control of blood glucose and mitigates the risk of complications associated with diabetes [ [22] , [23] , [24] ]. Various approaches, including statistical models [ [25] , [26] , [27] ], machine learning algorithms [ [28] , [29] , [30] , [31] ], and artificial neural networks [ [32] , [33] , [34] , [35] , [36] , [37] , [38] , [39] ], have been explored for predicting blood glucose levels. Statistical models, which utilize mathematical equations to project future blood glucose levels based on past data, insulin doses, and other pertinent factors, are among the simplest and most prevalent methods.…”