Table 2 Optimal performance achieved by each model using optimised architectures. Accuracy, precision, recall, and F1 score values are presented, and recurrent models (GRU, LSTM, BiGRU, BiLSTM) and the ARIMA model are compared.

From: Deep learning-based classification of hemiplegia and diplegia in cerebral palsy using postural control analysis

Model

Arch.

Accuracy

Precision

Recall

F1 Score

GRU

1

0.67

0.78

0.67

0.57

LSTM

1

0.71

0.80

0.71

0.66

Bi GRU

4

0.76

0.77

0.76

0.75

Bi LSTM

3

0.62

0.58

0.62

0.54

ARIMA

-

0.43

0.77

0.43

0.31

Back to article page