{"url":"/sota/traffic-prediction-on-metr-la","task":{"name":"Traffic Prediction","url":"/task/traffic-prediction","note":null},"dataset":{"name":"METR-LA","url":"/dataset/metr-la"},"category":"Time Series","categories":["Time Series"],"category_note":null,"description":"**Traffic Prediction** is a task that involves forecasting traffic conditions, such as the volume of vehicles and travel time, in a specific area or along a particular road. This task is important for optimizing transportation systems and reducing traffic congestion.\r\n\r\n<span style=\"color:grey; opacity: 0.6\">( Image credit: [BaiduTraffic](https://github.com/JingqingZ/BaiduTraffic) )</span>","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["MAE @ 12 step","12 steps MAE","12 steps MAPE","12 steps RMSE","MAE @ 3 step"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"MAE @ 12 step":"lower","12 steps MAE":"lower","12 steps MAPE":null,"12 steps RMSE":"lower","MAE @ 3 step":"lower"}},"counts":{"rows":20,"rows_with_code":17,"rows_with_paper_page":20,"rows_dated":20,"rows_using_additional_data":1},"rows":[{"rank_in_archive_order":1,"model":"TITAN","metrics":{"12 steps MAE":"3.08","12 steps MAPE":"8.43","12 steps RMSE":"6.21","MAE @ 12 step":"3.08","MAE @ 3 step":"2.41"},"uses_additional_data":false,"paper_date":"2024-09-26","paper":"/paper/a-time-series-is-worth-five-experts-1","paper_url":"https://arxiv.org/abs/2409.17440v1","paper_title":"A Time Series is Worth Five Experts: Heterogeneous Mixture of Experts for Traffic Flow Prediction","code":"https://github.com/sqlcow/TITAN","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"T-Graphormer","metrics":{"12 steps MAE":"3.19","12 steps MAPE":"8.62","12 steps RMSE":"6.12","MAE @ 12 step":"3.19","MAE @ 3 step":"2.63"},"uses_additional_data":false,"paper_date":"2025-01-22","paper":"/paper/t-graphormer-using-transformers-for","paper_url":"https://arxiv.org/abs/2501.13274v3","paper_title":"T-Graphormer: Using Transformers for Spatiotemporal Forecasting","code":"https://github.com/rdh1115/T-Graphormer","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"STGM","metrics":{"12 steps MAE":"3.229","12 steps MAPE":"9.39","12 steps RMSE":"7.099","MAE @ 12 step":"3.229"},"uses_additional_data":false,"paper_date":"2023-10-15","paper":"/paper/spatio-temporal-graph-mixformer-for-traffic","paper_url":"https://doi.org/10.1016/j.eswa.2023.120281","paper_title":"Spatio-Temporal Graph Mixformer for Traffic Forecasting","code":"https://github.com/Mouradost/STGM","n_code_links":1,"syntology":null},{"rank_in_archive_order":4,"model":"Traffic Transformer","metrics":{"MAE @ 12 step":"3.28"},"uses_additional_data":false,"paper_date":"2020-06-11","paper":"/paper/traffic-transformer-capturing-the-continuity","paper_url":"https://onlinelibrary.wiley.com/doi/abs/10.1111/tgis.12644","paper_title":"Traffic transformer: Capturing the continuity and periodicity of time series for traffic forecasting","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":5,"model":"SLCNN","metrics":{"MAE @ 12 step":"3.3"},"uses_additional_data":false,"paper_date":"2020-04-03","paper":"/paper/spatio-temporal-graph-structure-learning-for","paper_url":"https://aaai.org/ojs/index.php/AAAI/article/view/5470","paper_title":"Spatio-Temporal Graph Structure Learning for Traffic Forecasting","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":6,"model":"STAEformer","metrics":{"MAE @ 12 step":"3.34","MAE @ 3 step":"2.65"},"uses_additional_data":false,"paper_date":"2023-08-21","paper":"/paper/spatio-temporal-adaptive-embedding-makes","paper_url":"https://arxiv.org/abs/2308.10425v5","paper_title":"STAEformer: Spatio-Temporal Adaptive Embedding Makes Vanilla Transformer SOTA for Traffic Forecasting","code":"https://github.com/xdzhelheim/staeformer","n_code_links":1,"syntology":null},{"rank_in_archive_order":7,"model":"D2STGNN","metrics":{"MAE @ 12 step":"3.35","MAE @ 3 step":"2.56"},"uses_additional_data":false,"paper_date":"2022-06-18","paper":"/paper/decoupled-dynamic-spatial-temporal-graph","paper_url":"https://arxiv.org/abs/2206.09112v4","paper_title":"Decoupled Dynamic Spatial-Temporal Graph Neural Network for Traffic Forecasting","code":"https://github.com/zezhishao/d2stgnn","n_code_links":1,"syntology":null},{"rank_in_archive_order":8,"model":"STEP","metrics":{"MAE @ 12 step":"3.37","MAE @ 3 step":"2.61"},"uses_additional_data":false,"paper_date":"2022-06-18","paper":"/paper/pre-training-enhanced-spatial-temporal-graph","paper_url":"https://arxiv.org/abs/2206.09113v2","paper_title":"Pre-training Enhanced Spatial-temporal Graph Neural Network for Multivariate Time Series Forecasting","code":"https://github.com/zezhishao/step","n_code_links":2,"syntology":null},{"rank_in_archive_order":9,"model":"MegaCRN","metrics":{"MAE @ 12 step":"3.38","MAE @ 3 step":"2.63"},"uses_additional_data":false,"paper_date":"2022-11-27","paper":"/paper/spatio-temporal-meta-graph-learning-for","paper_url":"https://arxiv.org/abs/2211.14701v4","paper_title":"Spatio-Temporal Meta-Graph Learning for Traffic Forecasting","code":"https://github.com/deepkashiwa20/megacrn","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":2,"n_samples":2,"n_pointer_only_licence":1}},{"rank_in_archive_order":10,"model":"STD-MAE","metrics":{"12 steps MAE":"3.40","12 steps MAPE":"9.59","12 steps RMSE":"7.07","MAE @ 12 step":"3.40","MAE @ 3 step":"2.62"},"uses_additional_data":true,"paper_date":"2023-12-01","paper":"/paper/spatio-temporal-decoupled-masked-pre-training","paper_url":"https://arxiv.org/abs/2312.00516v3","paper_title":"Spatial-Temporal-Decoupled Masked Pre-training for Spatiotemporal Forecasting","code":"https://github.com/jimmy-7664/std-mae","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":3}},{"rank_in_archive_order":11,"model":"RGDAN","metrics":{"MAE @ 12 step":"3.40","MAE @ 3 step":"2.69"},"uses_additional_data":false,"paper_date":"2024-01-16","paper":"/paper/rgdan-a-random-graph-diffusion-attention","paper_url":"https://doi.org/10.1016/j.neunet.2023.106093","paper_title":"RGDAN: A random graph diffusion attention network for traffic prediction","code":"https://github.com/wengwenchao123/RGDAN","n_code_links":1,"syntology":null},{"rank_in_archive_order":12,"model":"ADN","metrics":{"MAE @ 12 step":"3.42"},"uses_additional_data":false,"paper_date":"2022-02-07","paper":"/paper/structured-time-series-prediction-without","paper_url":"https://arxiv.org/abs/2202.03539v1","paper_title":"Structured Time Series Prediction without Structural Prior","code":"https://github.com/radandreicristian/adn","n_code_links":1,"syntology":null},{"rank_in_archive_order":13,"model":"STAWnet","metrics":{"MAE @ 12 step":"3.44","MAE @ 3 step":"2.70"},"uses_additional_data":false,"paper_date":"2021-03-02","paper":"/paper/spatial-temporal-attention-wavenet-a-deep","paper_url":"https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/itr2.12044","paper_title":"Spatial‐temporal attention wavenet: A deep learning framework for traffic prediction considering spatial‐temporal dependencies","code":"https://github.com/CYBruce/STAWnet","n_code_links":1,"syntology":null},{"rank_in_archive_order":14,"model":"Finetune from t1-6 checkpoint","metrics":{"MAE @ 12 step":"3.47"},"uses_additional_data":false,"paper_date":"2019-12-11","paper":"/paper/incrementally-improving-graph-wavenet","paper_url":"https://arxiv.org/abs/1912.07390v1","paper_title":"Incrementally Improving Graph WaveNet Performance on Traffic Prediction","code":"https://github.com/nnzhan/Graph-WaveNet","n_code_links":4,"syntology":null},{"rank_in_archive_order":15,"model":"DCGCN","metrics":{"12 steps MAE":"3.48","12 steps MAPE":"9.94","12 steps RMSE":"6.94","MAE @ 12 step":"3.48"},"uses_additional_data":false,"paper_date":"2023-06-12","paper":"/paper/dynamic-causal-graph-convolutional-network","paper_url":"https://arxiv.org/abs/2306.07019v2","paper_title":"Dynamic Causal Graph Convolutional Network for Traffic Prediction","code":"https://github.com/MonBG/DCGCN","n_code_links":1,"syntology":null},{"rank_in_archive_order":16,"model":"GWNET-Cov","metrics":{"MAE @ 12 step":"3.50","MAE @ 3 step":"2.69"},"uses_additional_data":false,"paper_date":"2021-06-22","paper":"/paper/conditional-temporal-neural-processes-with","paper_url":"http://proceedings.mlr.press/v139/yoo21b.html","paper_title":"Conditional Temporal Neural Processes with Covariance Loss","code":"https://github.com/boseon-ai/Conditional-Temporal-Neural-Processes-with-Covariance-Loss","n_code_links":1,"syntology":null},{"rank_in_archive_order":17,"model":"Graph WaveNet","metrics":{"MAE @ 12 step":"3.53","MAE @ 3 step":"2.69"},"uses_additional_data":false,"paper_date":"2019-05-31","paper":"/paper/190600121","paper_url":"https://arxiv.org/abs/1906.00121v1","paper_title":"Graph WaveNet for Deep Spatial-Temporal Graph Modeling","code":"https://github.com/nnzhan/Graph-WaveNet","n_code_links":9,"syntology":{"n_ran":5,"n_unverified":2,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":18,"model":"ST-UNet","metrics":{"MAE @ 12 step":"3.55"},"uses_additional_data":false,"paper_date":"2019-03-13","paper":"/paper/st-unet-a-spatio-temporal-u-network-for-graph","paper_url":"https://arxiv.org/abs/1903.05631v2","paper_title":"ST-UNet: A Spatio-Temporal U-Network for Graph-structured Time Series Modeling","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":19,"model":"DCRNN","metrics":{"MAE @ 12 step":"3.6","MAE @ 3 step":"2.77"},"uses_additional_data":false,"paper_date":"2017-07-06","paper":"/paper/diffusion-convolutional-recurrent-neural","paper_url":"http://arxiv.org/abs/1707.01926v3","paper_title":"Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting","code":"https://github.com/dmlc/dgl/tree/master/examples/pytorch/dtgrnn","n_code_links":19,"syntology":{"n_ran":17,"n_unverified":18,"n_samples":35,"n_pointer_only_licence":15}},{"rank_in_archive_order":20,"model":"STGCN","metrics":{"MAE @ 12 step":"4.45"},"uses_additional_data":false,"paper_date":"2017-09-14","paper":"/paper/spatio-temporal-graph-convolutional-networks","paper_url":"http://arxiv.org/abs/1709.04875v4","paper_title":"Spatio-Temporal Graph Convolutional Networks: A Deep Learning Framework for Traffic Forecasting","code":"https://github.com/VeritasYin/STGCN_IJCAI-18","n_code_links":7,"syntology":{"n_ran":3,"n_unverified":0,"n_samples":3,"n_pointer_only_licence":3}}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 7,081 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9623,"papers_checked":7081,"papers_extracted_not_yet_verified":217,"boards_without_verdict":27,"papers_not_yet_extracted":2325},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":5,"rows_with_any_sample_ran":4,"distinct_papers_with_graph_line":5,"distinct_papers_with_any_sample_ran":4,"samples_over_distinct_papers":{"n_ran":27,"n_unverified":23,"n_samples":50,"n_pointer_only_licence":22,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":27,"n_unverified":23,"n_samples":50,"n_pointer_only_licence":22,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}