Add to Favourites
To login click here
This article examines the performance of various algorithms in estimating sleep outcomes, such as wake after sleep onset (WASO). The Oakley-rescore algorithm showed the lowest root mean squared error (RMSE) and narrowest confidence interval in estimating WASO as compared to the PSG-derived WASO. Other algorithms such as Cole-Kripke and van Hees also presented low RMSE scores and narrow confidence intervals. Machine learning models such as Random Forest, LSTM-50 and LSTM-100 also showed low mean errors, but high RMSE scores, wide confidence intervals and low correlation with PSG-derived WASO.
