The bradycardia EWS (early warning system) binary classification/prediction model uses a hybrid CNN/RNN system. The CNN extracts features from the data, and the large RNN (bidirectional LSTM) captures long-term dependencies inside the data. Trained off alonso130r/bradycardia-balanced (augmented to have a 50/50 ratio of bradycardia to no bradycardia) and evaluated on alonso130r/HTR-2024-brachy.
