Table 7.

Relative ranks of ML^a models for predicting BG^b levels in PH^c=45 minutes.

ML model	SUCRAd	Relative rank
Convolutional recurrent neural network multitask learning (CRNN-MTL)	52.1	5.8
Convolutional recurrent neural network multitask learning glycemic variability (CRNN-MTL-GV)	41.8	6.8
Convolutional recurrent neural network transfer learning (CRNN-TL)	31.6	7.8
Convolutional recurrent neural network single-task learning (CRNN-STL)	27.5	8.2
SVMe	32.0	7.8
k-Nearest neighbor (kNN)	61.4	4.9
DTf	26.3	8.4
RFg	70.3	4.0
AdaBoost	34.1	7.6
XGBoosth	73.5	3.7
NNMi	99.4	1.1

^aML: machine learning.

^bBG: blood glucose.

^cPH: prediction horizon.

^dSUCRA: surface under the cumulative ranking.

^eSVM: support vector machine.

^fDT: decision tree.

^gRF: random forest.

^hXGBoost: Extreme Gradient Boosting.

ⁱNNM: neural network model.