{"title":"Graph Multi-Attention Network-based Taxi Demand Prediction","authors":"Haifan Tang, Youkai Wu, Zhaoxia Guo","doi":"10.1109/DOCS55193.2022.9967748","DOIUrl":null,"url":null,"abstract":"Taxi is an important component of the urban transport system in most cities. Accurate taxi demand prediction can effectively reduce the waiting time of passengers and shorten the no-load travel of drivers, which is helpful in alleviating traffic congestion and improving traffic efficiency. Due to the complexity of the traffic system and spatiotemporal dependencies among regions in a road network, traditional prediction methods cannot predict taxi demands of different regions effectively. This paper introduces a Graph Multi-Attention Network (GMAN) to handle the taxi demand prediction problem with better performance, which aims to predict the taxi demands in all regions of a road network in the next time period. The effectiveness of the GMAN is validated based on a large-scale dataset of taxi demands from a real urban road network. Experimental results show that the GMAN outperforms 5 commonly used benchmarking models, including 3 state-of-the-art machine learning models.","PeriodicalId":348545,"journal":{"name":"2022 4th International Conference on Data-driven Optimization of Complex Systems (DOCS)","volume":"64 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 4th International Conference on Data-driven Optimization of Complex Systems (DOCS)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/DOCS55193.2022.9967748","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Taxi is an important component of the urban transport system in most cities. Accurate taxi demand prediction can effectively reduce the waiting time of passengers and shorten the no-load travel of drivers, which is helpful in alleviating traffic congestion and improving traffic efficiency. Due to the complexity of the traffic system and spatiotemporal dependencies among regions in a road network, traditional prediction methods cannot predict taxi demands of different regions effectively. This paper introduces a Graph Multi-Attention Network (GMAN) to handle the taxi demand prediction problem with better performance, which aims to predict the taxi demands in all regions of a road network in the next time period. The effectiveness of the GMAN is validated based on a large-scale dataset of taxi demands from a real urban road network. Experimental results show that the GMAN outperforms 5 commonly used benchmarking models, including 3 state-of-the-art machine learning models.