Recognition of Transportation State by Smartphone Sensors Using Deep Bi-LSTM Neural Network

Table 2

Experiment configuration.

Parameter	Value
Size of input vector	768
Size of batch	128
LRB (learning rate base)	0.5–0.005
LRD (learning rate decay)	0.90–0.95
Activation function	ReLU
RE (regularization rate)	0.5–0.005
HP (hidden points)	50–300
MAD (moving average decay)	0.99
Epochs	1000–1400
Probability of dropout	0.5–1