Traffic Prediction On Metr La

Metrics

MAE @ 12 step

Results

Performance results of various models on this benchmark

Model Name	MAE @ 12 step	Paper Title	Repository
SLCNN	3.3	Spatio-Temporal Graph Structure Learning for Traffic Forecasting	-
T-Graphormer	3.19	T-Graphormer: Using Transformers for Spatiotemporal Forecasting
MegaCRN	3.38	Spatio-Temporal Meta-Graph Learning for Traffic Forecasting
STGM	3.229	Spatio-Temporal Graph Mixformer for Traffic Forecasting
TITAN	3.08	A Time Series is Worth Five Experts: Heterogeneous Mixture of Experts for Traffic Flow Prediction
Traffic Transformer	3.28	Traffic transformer: Capturing the continuity and periodicity of time series for traffic forecasting	-
STGCN	4.45	Spatio-Temporal Graph Convolutional Networks: A Deep Learning Framework for Traffic Forecasting
STEP	3.37	Pre-training Enhanced Spatial-temporal Graph Neural Network for Multivariate Time Series Forecasting
GWNET-Cov	3.50	Conditional Temporal Neural Processes with Covariance Loss	-
DCGCN	3.48	Dynamic Causal Graph Convolutional Network for Traffic Prediction
DCRNN	3.6	Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting
STD-MAE	3.40	Spatial-Temporal-Decoupled Masked Pre-training for Spatiotemporal Forecasting
Graph WaveNet	3.53	Graph WaveNet for Deep Spatial-Temporal Graph Modeling
ST-UNet	3.55	ST-UNet: A Spatio-Temporal U-Network for Graph-structured Time Series Modeling	-
ADN	3.42	Structured Time Series Prediction without Structural Prior
STAEformer	3.34	STAEformer: Spatio-Temporal Adaptive Embedding Makes Vanilla Transformer SOTA for Traffic Forecasting
RGDAN	3.40	RGDAN: A random graph diffusion attention network for traffic prediction
Finetune from t1-6 checkpoint	3.47	Incrementally Improving Graph WaveNet Performance on Traffic Prediction
D2STGNN	3.35	Decoupled Dynamic Spatial-Temporal Graph Neural Network for Traffic Forecasting
STAWnet	3.44	Spatial‐temporal attention wavenet: A deep learning framework for traffic prediction considering spatial‐temporal dependencies

0 of 20 row(s) selected.