OpenStreetMap based POI knowledge graph enhanced by large language model
文献类型：期刊论文
| 作者 | Zhang, Yifan4; Chen, Yizhe3; Wang, Zhiyun3; Song, Ci1,2; Yu, Wenhao3 |
| 刊名 | JOURNAL OF GEOGRAPHICAL SYSTEMS
 |
| 出版日期 | 2026-04-21 |
| 卷号 | N/A |
| 关键词 | POI Knowledge graph Large language model OpenStreetMap |
| ISSN号 | 1435-5930 |
| DOI | 10.1007/s10109-026-00494-7 |
| 产权排序 | 3 |
| 文献子类 | Article ; Early Access |
| 英文摘要 | Point-of-interest knowledge graphs (PKGs) based on OpenStreetMap (OSM) play a significant role in various domains, such as smart cities, navigation, and personalized recommendation systems. However, OSM data suffers from significant sparsity issues. Although existing PKG construction methods alleviate this problem to some extent by incorporating multi-source datasets, they typically focus on certain POI attributes and overlook fine-grained attributes and implicit relationships between POIs. Moreover, methods that enhance PKGs using multi-source datasets are costly and lack universality. To address these issues, this study proposes an enhancement of OSM-based PKGs using Large Language Models (LLMs). LLMs, pretrained with vast amounts of factual and relational knowledge, serve as an external knowledge base to enrich the semantic information in PKGs. We validate this method using 54,797 POI data entries from Wuhan city on OSM, and experimental results show that the LLM-enhanced PKG adds 14 key attribute types and 4 implicit relationship types, with significant improvements in attribute coverage and performance in the question-answering system. This study not only provides an effective solution to the sparsity of OSM data but also offers new practical insights for PKG construction in multi-task application scenarios. |
| URL标识 | 查看原文 |
| WOS研究方向 | Geography |
| 语种 | 英语 |
| WOS记录号 | WOS:001745320500001 |
| 出版者 | SPRINGER HEIDELBERG |
| 源URL | [http://ir.igsnrr.ac.cn/handle/311030/221473]  |
| 专题 | 资源与环境信息系统国家重点实验室_外文论文 |
| 通讯作者 | Yu, Wenhao |
| 作者单位 | 1.Univ Chinese Acad Sci, Coll Resources & Environm, Beijing, Peoples R China 2.Chinese Acad Sci, Inst Geog Sci & Nat Resources Res, State Key Lab Resources & Environm Informat Syst, Beijing, Peoples R China; 3.China Univ Geosci, Sch Geog & Informat Engn, Wuhan, Peoples R China; 4.China Univ Min & Technol, Sch Environm Sci & Spatial Informat, Xuzhou, Peoples R China; |
| 推荐引用方式 GB/T 7714 | Zhang, Yifan,Chen, Yizhe,Wang, Zhiyun,et al. OpenStreetMap based POI knowledge graph enhanced by large language model[J]. JOURNAL OF GEOGRAPHICAL SYSTEMS,2026,N/A. |
| APA | Zhang, Yifan,Chen, Yizhe,Wang, Zhiyun,Song, Ci,&Yu, Wenhao.(2026).OpenStreetMap based POI knowledge graph enhanced by large language model.JOURNAL OF GEOGRAPHICAL SYSTEMS,N/A. |
| MLA | Zhang, Yifan,et al."OpenStreetMap based POI knowledge graph enhanced by large language model".JOURNAL OF GEOGRAPHICAL SYSTEMS N/A(2026). |
入库方式： OAI收割
来源：地理科学与资源研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。