. 2022 Feb 8;25(3):103888. doi: 10.1016/j.isci.2022.103888

Table 3.

Comparing the effect of aerobic fitness on the performance of models designed to predict exercise-related changes in glucose

Model	Personalization	Accuracy [%]		MARE [%]
Model	Personalization	Lower VO₂max n = 70 obs	Higher VO₂max n = 88 obs	Lower VO₂max n = 70 obs	Higher VO₂max n = 88 obs
Predicting minimum glucose during exercise

MARS Model	Population	73	65	23	20
MARS Model	Adaptation	84	72	20	20
MARS Model + Exercise History	Population	74	65	19	19
MARS Model + Exercise History	Adaptation	86	70	19	20
AR Model	Population	88	75^a	16	38^a
AR Model	Adaptation	85	82	16	34^a
Logistic Regression	Population	65	55	–	–
Logistic Regression	Adaptation	72	65	–	–

Predicting minimum glucose in the 4 hours following exercise

MARS Model	Population	57	55	25	21
MARS Model	Adaptation	70	69	22	20
MARS Model + Exercise History	Population	63	52	25	24
MARS Model + Exercise History	Adaptation	69	68	23	20^a
Logistic Regression	Population	58	56	–	–
Logistic Regression	Adaptation	72	63	–	–

indicates that the significance p < 0.05 as determined by Wilcoxon rank-sum test for nonparametric data, comparing algorithm performance on participants with higher aerobic and lower aerobic fitness rankings.