TABLE VIII.

Training and validation performance for experiment 1 with all values in mean percentages and 95% confidence intervals

	MGH training (binary)	MGH validation (binary)		MGH training (multiclass)	MGH validation (multiclass)
Accuracy	95.8 [95.8–95.9]	95.9 [95.8–95.9]	Accuracy	95.8	99.1
Sensitivity	67.2 [67.0–67.3]	69.2 [68.8–69.5]	Sensitivity	67.2	50.0
Specificity	97.8 [97.8–97.8]	97.6 [97.6–97.7]	Specificity	97.8	99.5
Precision	66.9 [66.7–67.0]	66.0 [65.6–66.3]	Precision	66.9	36.9
F1-score	67.0 [66.9–67.1]	67.5 [67.2–67.8]	F1-score	67.0	40.4
Cohen’s kappa	64.8 [64.7–64.9]	65.3 [65.0–65.6]	Cohen’s kappa	37.2 [37.1–37.3]	36.8 [36.6–37.1]
Multi-class training	Sensitivity	Specificity	Multi-class validation	Sensitivity	Specificity
Obstructive apnea	52.4 [52.1–52.6]	99.7 [99.7–99.7]	Obstructive apnea	48.4 [48.1–49.5]	99.7 [99.7–99.7]
Central apnea	84.9 [84.7–85.1]	99.6 [99.6–99.6]	Central apnea	87.6 [87.0–88.2]	99.6 [99.5–99.6]
RERA	40.9 [40.7–41.1]	98.9 [98.9–98.9]	RERA	40.6 [40.0–41.2]	99.0 [99.0–99.0]
Hypopnea	23.2 [23.1–23.4]	99.7 [99.7–99.7]	Hypopnea	22.9 [22.4–23.3]	99.7 [99.7–99.7]
Mean	50.4	99.5	Mean	50.0	99.5
Multi-class training	Precision	F1-score	Multi-class validation	Precision	F1-score
Obstructive apnea	45.2 [44.9–45.4]	48.5 [48.3–48.7]	Obstructive apnea	42.0 [41.3–42.7]	45.1 [44.5–45.7]
Central apnea	42.5 [42.3–42.7]	56.6 [56.4–56.8]	Central apnea	41.7 [41.1–42.2]	56.5 [55.9–57.1]
RERA	24.4 [24.3–24.5]	30.5 [30.4–30.7]	RERA	25.5 [25.1–25.8]	31.3 [30.9–31.7]
Hypopnea	39.9 [39.7–40.2]	29.4 [29.2–29.5]	Hypopnea	38.5 [37.8–39.1]	28.7 [28.2–29.2]
Mean	38.0	41.3	Mean	36.9	40.4