From: A deep learning framework for predicting cyber attacks rates
W x | Weight matrix connecting the input layer and the hidden layer |
W h | Weight matrix connecting two consecutive hidden states |
W y | Weight matrix connecting the hidden state and the output layer |
b h | Bias vector in hidden layer |
b y | Bias vector in output layer |
h t | Hidden state at time t |
σ(·) | Activation function |
x t | Input at time t |
y t | Real output at time t |
\(\hat {y}_{t}\) | Predicted output at time t |
J | Objective function |