摘要
蕴含静态路网特性和动态交通特性的轨迹大数据具有多源异构的特点,给交通知识分析和挖掘带来了困难。近年兴起的知识图谱可对多源异构数据进行有效融合、对齐和加工。基于轨迹大数据,采用自底向上的方法在结构化数据中构建司机出行知识图谱,还原轨迹数据的时空关系和语义关联。具体来说,通过RDF Mapping进行知识抽取,将结构化数据映射为RDF数据;再利用Neo4j存储对静态路网拓扑进行知识补齐和融合。司机出行知识图谱的构建为交通出行知识查询和推理奠定了良好基础;同时,从结构化数据中自底而上构建知识图谱的方法可被扩展应用于其他领域的结构化数据,对推广知识图谱应用起到推进作用。
The trajectory big data containing static road network characteristics and dynamic traffic characteristics has the characteristics of multi-source heterogeneous,which brings some difficulties to traffic knowledge analysis and mining.In recent years,knowledge graph is an effective means to fuse,align and process multi-source heterogeneous data.Based on the trajectory big data,we realized the construction of bottom-up driver travel knowledge graph from the structured data,and restored the spatio-temporal relationship and semantic association in the trajectory data.Specifically,we used RDF Mapping to extract knowledge and map structured data to RDF data.Then,we stored the extracted knowledge in Neo4j,complemented the new knowledge,and integrated it into the static road network topology.The effective construction of driver travel knowledge graph has laid a good foundation for traffic travel knowledge query and reasoning.At the same time,the explored method of constructing the knowledge graph from the structured data can be extended to the structured data in other fields,which promotes the application of the knowledge graph.
作者
李璇
吴雷
姜梦
潘晓
LI Xuan;WU Lei;JIANG Meng;PAN Xiao(Shijiazhuang Tiedao University,Shijiazhuang 050043,China;School of Information Science and Engineering,Yanshan University,Qinhuangdao 066004,China)
出处
《地理空间信息》
2023年第1期32-35,130,共5页
Geospatial Information
基金
国家自然科学基金资助项目(61472340,61303017)
河北省自然科学基金资助项目(F2021210005)
河北省省级科技计划资助创新能力提升计划资助项目(21550803D)
河北省教育厅青年拔尖资助项目(BJ2021085)
河北省科技厅大学、中学在校生科技创新能力培育专项资助项目(DXS202106,22E50118D)
国家级大学生创新创业训练计划资助项目(202110107015)。
