STGNN-FAM: A Traffic Flow Prediction Model for Spatiotemporal Graph Networks Based on Fusion of Attention Mechanisms

Table 1

STGNN-FAM model prediction algorithm.

---

Input: historical traffic flow speed data , historical time step P, forecast time step Q

Output: Predicted traffic flow speed data

(1) Use the Speed2Vec mechanism to process the historical traffic flow speed data into a form that can be sent to the GAT network for calculation

(2) function GAT()

(3)  for i = 1, …, N − 1, N do

(4)   

(5)   

(6)  end for

(7) 

(8) return

(9) end function

(10) Reshape feature output by GAT network back to

(11) for j = P − 1, …, 1, 0 do

(12)   If j = P − 1

(13)    

(14)    

(15)  end for

(16) 

(17) function Attention (H)

(18)  

(19)  

(20) return z

(21) end function

(22) 

(23) return

---