Multicenter Validation of an AI-Based Snoring Model vs. Wearables for Diagnosing Sleep Apnea.

Snoring-Sound–Based Model Shows Comparable Performance with Numerically Higher AUCs When directly compared on paired samples using DeLong’s test ( Figure 5 ), the snore-based model generally achieved higher AUCs than the wearable device-based model across most PSG-defined classes, although the differences did not reach statistical significance (all p > 0.05).

also in 111,027 other papers