Attention Mechanism with Spatial-Temporal Joint Deep Learning Model for the Forecasting of Short-Term Passenger Flow Distribution at the Railway Station

Algorithm 1

.

	Input: Directed nonweight graph
	window size
	embedding size
	walks per set vertex
	walk length
	Output: Vertex representation matrix
(1)	Initialization: Sliding window sampling from
(2)	Construct a binary Tree from
(3)	for to do
(4)	= Shuffle ()
(5)	for each do
(6)
(7)
(8)	end for
(9)	end for