Fig. 7
From: Detecting epileptic seizures with electroencephalogram via a context-learning model
The framework of the EEG context learning algorithm. This framework adopts a neural network structure to extract the temporal features from EEG contexts. The input layer is a projection layer that projects each word to an unique vector. The hidden layer is an integration of the vectors in a context. And the output layer is a classifier where each tuple in the output corresponds to the conditional probability of a word in the dictionary