Table 4 The optimal combination of hyperparameters.

From: Empowering stroke recovery with upper limb rehabilitation monitoring using TinyML based heterogeneous classifiers

Considerations

Optimal value

Training set

80%

Test set

20%

Input layer

126 features

Dense layer

40 neurons

Dense layer

20 neurons

Dropout

rate 0.25

Number of training cycles

60

Learning rate

0.0005

Back to article page