Fig. 5. Impact of training dataset size and recency on discharge prediction performance in the test dataset (01 February 2019 to 31 January 2020).

a By increasing training data size. b By decreasing recency of same-size training data. 95% confidence intervals were calculated using bootstrapping (n = 500). PPV: positive predictive value, NPV: negative predictive value, MAE: normalised mean absolute error (mean difference in predicted and actual discharges per day divided by the mean number of discharges per day).