[GNN4TRAFFIC]2018 - T-GCN- A Temporal Graph Convolutional Network for Traffic Prediction

Background

Considering the Spatial dependency in non-euclidean structure

Datasets:

the taxi speed data of the Luohu District in

• Shenzhen datasets
• Los-loop datasets

Graph construction:

• The Road graph G: an unweighted graph $G=(V,E)$, $V$ is a set of road nodes, $V=v1,v2,\cdot \cdot \cdot ,vN$, $N$ is the number of the nodes, and $E$ is a set of edges.

• The adjacency matrix $A$: represent the connection between roads.(0/1)

• The feature matrix $X:X\in R^{N\times P}$ represents the number of node attribute features (the length of the historical time series) and $X_t\in R^{N\times i}$ is used to represent the speed on each road at time $i$

Problem Definition:

$[X_{t+1},\cdot \cdot \cdot ,X_{t+T}]=f(G;(X_{t-n},\cdot \cdot \cdot ,X_{t-1},X_t))$

Model:

1. Statistic spatial feature -> 2-LAYER GCN:

   $f(X,A)=\sigma (\hat{A}Relu(\hat{A}X{W}_0)W_1)$

   where, $\hat{A}={\tilde{D}}^{-\frac{1}{2}}\tilde{A}{\tilde{D}}^{-\frac{1}{2}}$ denotes preprocessing step, $\tilde{A}=A+I_N$ is a matrix with self-connection structure , $\tilde{D}={\sum }_j{\tilde{A}}_{ij}$ is a degree matrix, $W_0$ and $W_1$ represent the weight matrix in the first and second layer, and $\sigma (\cdot )$, $Relu()$ represent the activation function

2. GRU
   u_t = \sigma(W_u[f(A, X_t), h_{t-1} + b_u)
   r_t = \sigma(W_r[f(A, X_t), h_{t-1} + b_r)
   c = tanh(W_c[f(A, X_t), (r_t * h_{t-1})]) + b_c)
   h_t = u_t *h_{t-1} + (1-u_t)*c_t

3. Loss Function :
   $loss=\Vert Y_t-{\hat{Y}}_t\Vert +\lambda L_{reg}$
   where $L_{reg}$ is an $L2$ regularization term that helps to avoid an over fitting problem and $\lambda$ is a hyperparameter

Model Parameters:

• Learning rate:0.001
• Batch size:64
• Training epoch: 3000
• Optimizer:the Adam optimizer.
• The number of hidden layers: select from [8, 16, 32, 64, 100, 128], comparing different evaluation matrix based on different hidden layers, and then select the best one

Perturbation Analysis and Robustness:

• Add two types of commonly random noise to the data during the experiment
• The random noise obeys the Gaussian distribution $N\in (0,\sigma 2)(\sigma \in (0.2,0.4,0.8,1,2))$ & the Poisson distribution $P(\lambda )(\lambda \in (1,2,4,8,16))$ and then normalized the values of the noise matrices turn to [0, 1]

Further discuss:

• Predict poorly at the peak
• Certain errors between the real traffic information and the prediction results: no record when no taxis on the roads
• But it can detect the start and end of the rush hour and
  make prediction results with similar pattern with the real traffic speed

2018 - T-GCN- ATemporal Graph Convolutional Network for Traffic Prediction¹

---

1. T-GCN- ATemporal Graph Convolutional Network for Traffic Prediction ↩︎