.png){kind=icon}
返回
BEKG: A built environment knowledge graph
.png){kind=icon}
OA
AI
DOI：10.1080/09613218.2023.2238851.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
In recent years, the digitalization of the built environment has progressed rapidly due to the development of modern design and construction technologies. However, the need for extensive professional knowledge in this field has not been met by practitioners and scholars. To address this problem, a study was conducted to build a knowledge graph in the built environment domain, which stores entities and their connections in a graph data model. To achieve it, this research collected more than 80,000 paper abstracts from the built environment domain. To ensure the accuracy of entities and relationships in the knowledge graph, two well-annotated datasets were created with 29 types of relationships, each containing 2000 and 1450 instances, respectively, for Named Entity Recognition (NER) and relationship extraction (RE) tasks. Two BERT-based models were trained on these datasets and achieved over 85% accuracy in both tasks. Using these models, over 200,000 high-quality relationships and entities were extracted from abstract data. This comprehensive knowledge graph will help practitioners and scholars better understand the built environment domain.
Keyword:
Built environment
knowledge graph
machine learning
natural language processing
scientific knowledge discovery
期刊
B
IF:
3.4
论文数:
1.6K
被引数:
4.2K
机构
引用论文
Kinematic analysis of limb movements in neuropsychological research: Subtle deficits and recovery of function.神经心理学研究中肢体运动的运动学分析: 细微的缺陷和功能的恢复。
.png){kind=icon}
PREAI
分享收藏Knowledge graph for identifying hazards on construction sites: Integrating computer vision with ontology用于识别建筑工地危险的知识图谱: 将计算机视觉与本体集成
OAAI
OAAI分享收藏
查看更多内容