Short-Term IoT Data Forecast of Urban Public Bicycle Based on the DBSCAN-TCN Model for Social Governance

Table 3

Parameter setting for the DBSCAN-TCN model.

Parameter	Setting
Activation functions	ReLu
Number of time steps
Number of hidden cells per layer	24
Input layer dimension
Output layer latitude
Number of cycles
Number of layers of the temporal convolutional network	4 layers
Convolutional kernel size	3
Dropout	0.5
Number of samples per batch
Learning rate