Fig. 7: The mRMEBP framework.
From: mRMEBP: a unified framework for online detection of atrial fibrillation utilizing deep learning
a The input layer with 5 features. b The hidden layer comprising 31 neurons and utilizing the sigmoid activation function. c The output from the hidden layer. d The Softmax layer for classification. e One-hot encoded representations of 6 categorical labels. f The distribution of class predictions for the input data set. See text for abbreviations.
