2026-10-14T15:59:51Z https://www.scidb.cn/oai

10.57760/sciencedb.37760 2026-08-03T16:11:59Z

2026-08-03 A Dataset for Constructing a Hyper-Structured Knowledge Graph from the Code for Geotechnical Investigations of Urban Rail Transit doi:10.57760/sciencedb.37760 en This dataset is derived from Chapters 5–9 of the Code for Geotechnical Investigations of Urban Rail Transit (GB 50307—2012). It covers technical content related to feasibility study investigation, preliminary investigation, detailed investigation, construction investigation, and geotechnical investigations for construction methods, and provides hyper-structured quintuple data for knowledge graph construction in urban rail transit geotechnical engineering investigation. The dataset is organized as one Excel workbook containing one primary data worksheet. It comprises 2,284 hyper-structured quintuple records covering 346 text paths, including individual provisions and their subordinate items. Each record contains seven fields: ID, Head Node, Head-Relation Edge, Relation Node, Relation-Tail Edge, Tail Node, and Text Path. After global deduplication and node consolidation, the dataset contains 668 content nodes, 600 modifier nodes, and 187 relation nodes. The seven edge types are Subject-Predicate Edge, Verb-Object Edge, Adverbial Edge, Adverbial-Center Edge, Attributive Edge, Attributive-Center Edge, and Property Edge. Relation nodes explicitly represent core actions, conditional relations, and modification relations in the code provisions, while the Text Path field maintains traceability between structured knowledge units and the original provisions. The dataset can be used for knowledge graph construction, Neo4j graph database import, semantic retrieval, source-text tracing, multi-hop association analysis, knowledge services, and intelligent question answering. The data file is provided in .xlsx format and can be opened using Microsoft Excel, WPS Office, or other compatible spreadsheet software. urban rail transit; geotechnical engineering investigation; standard code; quintuple; knowledge graph; hyper-structure Cai Wenkai Wang Shu Ji Tao Ye Peng PUBLIC https://creativecommons.org/licenses/by/4.0/ dataset Science Data Bank