Learning Spatial-Temporal Features of Ride-Hailing Services with Fusion Convolutional Networks

Algorithm 1

LC-ST-FCN training algorithm.

	Input: Historical demand of each region: ; lengths of input data sequence: ; length of the period interval: .
	Output: Learned LC-ST-FCN model.
(1)	//construct training instance:
(2)
(3)	for all the available time interval do
(4)
(5)
(6)	// is the target at time
(7)	put a training instance into
(8)	end for
(9)	//Training:
(10)	repeat
(11)	Initialize the biases and weights at each layer;
(12)	Sample minibatch from randomly;
(13)	if in 2D or 3D convolution layers then
(14)	for filters do
(15)	//Parameter sharing
(16)	optimize learnable parameters and
(17)	end for
(18)	end if
(19)	if in locally connected convolution layers then
(20)	for filters do
(21)	//Without parameter sharing
(22)	optimize learnable parameters and
(23)	end for
(24)	end if
(25)	Calculate the stochastic gradient by minimizing the objective function (4);
(26)	Update the parameters via backpropagation;
(27)	until stopping criteria are met