|
Algorithm 2 Pseudo code for the GA-based hybrid recurrent neural network (RNN) model. |
Begin
-
1.
C = 0
-
2.
initialize population P(c)
-
3.
RMSE from LSTM(c)
-
4.
RMSE from GRU(c)
-
5.
ensemble LSTM(c) & GRU(c) and compute fitness
-
6.
C = c + 1
-
7.
if termination criterion achieved go to step 12
-
8.
select p(c) from p(c+1)
-
9.
execute crossover p(c)
-
10.
perform mutatation p(c)
-
11.
go to step 3
-
12.
output optimal parameters
End = 0 |