Table 5 Hyperparameter choices made for best-performing models constructed by optimizers.

From: Tuning attention based long-short term memory neural networks for Parkinson’s disease detection using modified metaheuristics

Method

Learning rate

Dropout

Epochs

Layers

Neurons L1

Neurons L2

LSTM-ATT-MCOA

0.010000

0.050000

60

1

15

N/a

LSTM-ATT-COA

0.010000

0.059956

60

2

14

10

LSTM-ATT-GA

0.010000

0.180735

60

1

15

N/a

LSTM-ATT-PSO

0.010000

0.052466

60

1

15

N/a

LSTM-ATT-FA

0.009148

0.153722

44

1

12

N/a

LSTM-ATT-GWO

0.010000

0.200000

60

2

15

15

LSTM-ATT-BSO

0.010000

0.200000

60

1

15

N/a

LSTM-ATT-COLSHADE

0.008276

0.200000

60

1

15

N/a

Back to article page