Skip to main content
. 2020 Aug 22;9(9):2718. doi: 10.3390/jcm9092718

Figure 2.

Figure 2

Machine learning architecture of the model used in the study. Hybrid neural network constructed of a recurrent neural network (Long Short-term Memory followed by dropout layer) and convolutional neural network (convolutional, Rectified linear unit (ReLU) activation unit and dropout and global pooling layers). Inputs were variables listed, with outputs being neurological injury classification.