{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/revisiting-spatial-temporal-similarity-a-deep","title":"Revisiting Spatial-Temporal Similarity: A Deep Learning Framework for Traffic Prediction","arxiv_id":"1803.01254","date":"2018-03-03","proceeding":null,"authors":["Huaxiu Yao","Xianfeng Tang","Hua Wei","Guanjie Zheng","Zhenhui Li"],"abstract":"Traffic prediction has drawn increasing attention in AI research field due to\nthe increasing availability of large-scale traffic data and its importance in\nthe real world. For example, an accurate taxi demand prediction can assist taxi\ncompanies in pre-allocating taxis. The key challenge of traffic prediction lies\nin how to model the complex spatial dependencies and temporal dynamics.\nAlthough both factors have been considered in modeling, existing works make\nstrong assumptions about spatial dependence and temporal dynamics, i.e.,\nspatial dependence is stationary in time, and temporal dynamics is strictly\nperiodical. However, in practice, the spatial dependence could be dynamic\n(i.e., changing from time to time), and the temporal dynamics could have some\nperturbation from one period to another period. In this paper, we make two\nimportant observations: (1) the spatial dependencies between locations are\ndynamic; and (2) the temporal dependency follows daily and weekly pattern but\nit is not strictly periodic for its dynamic temporal shifting. To address these\ntwo issues, we propose a novel Spatial-Temporal Dynamic Network (STDN), in\nwhich a flow gating mechanism is introduced to learn the dynamic similarity\nbetween locations, and a periodically shifted attention mechanism is designed\nto handle long-term periodic temporal shifting. To the best of our knowledge,\nthis is the first work that tackles both issues in a unified framework. Our\nexperimental results on real-world traffic datasets verify the effectiveness of\nthe proposed method.","url_abs":"http://arxiv.org/abs/1803.01254v2","url_pdf":"http://arxiv.org/pdf/1803.01254v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"revisiting-spatial-temporal-similarity-a-deep","repo_url":"https://github.com/tangxianfeng/STDN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"revisiting-spatial-temporal-similarity-a-deep","repo_url":"https://github.com/aissahm/Forecasting_Citywide_Crowd_Flows","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"revisiting-spatial-temporal-similarity-a-deep","repo_url":"https://github.com/giallo41/taxi-demand","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"revisiting-spatial-temporal-similarity-a-deep","repo_url":"https://github.com/tpepin96/NYCDatasetProcessing","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"revisiting-spatial-temporal-similarity-a-deep","repo_url":"https://github.com/zzwells/jdd2018-population-forecast","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"traffic-prediction","task_name":"Traffic Prediction"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1803.01254","atlas_url":"https://app.syntology.ai/?focus=1803.01254","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}