Wrist EMG Monitoring Using Neural Networks Techniques

Table 1

Features of the BP-ANN model.

Feature	Description	References
Topology	MLP	[23]
Inputs	Vinp	—
Outputs	Vout	—
Hidden layer and neuron	one layer with 25 neurons	[24]
Activation function (input-hidden-output)	Fiden–Ftanh–Ftanh	[25]
Training algorithm	Extended back-propagation	—
Learning parameters	MSE simple	[21]
Cost function	MSE simple	—
Weight update	Batch	[22]
Weight initialization	Haykin heuristic	[26]
Convergence criteria	Epochs (1,000–200,000)	—