Skip to main content
Fig. 3 | BMC Medical Informatics and Decision Making

Fig. 3

From: Recursive neural networks in hospital bed occupancy forecasting

Fig. 3

Open loop training and closed loop prediction of NARX on an example sequence of bed occupation data with delay d = 2. Prediction and training is handled identically for training and evaluation data. Supporting data is not shown for simplicity, as it is always taken from pre-calculated time series (top bar) and never from prediction

Back to article page