CQ2Onto
收藏资源简介:
CQ2Onto是一个用于评估基于大型语言模型(LLM)辅助的从能力问题(Competency Questions, CQs)生成本体的基准数据集,涵盖六个不同领域(Wine、AWO、ODRL、Water、VGO和SWO)。数据集为每个领域提供黄金标准OWL本体、两个CQ文件(对应CQ2Term和CQ2Onto任务)以及注释表格。CQ2Term任务要求从单个CQ中提取类和属性,而CQ2Onto任务要求从一组CQs生成完整OWL本体。数据集文件包括JSON格式的CQ列表和RDF/XML格式的OWL本体文件,适用于本体工程、本体生成和术语提取等任务的研究和评估。
CQ2Onto is a benchmark dataset for evaluating Large Language Model (LLM)-assisted ontology generation from Competency Questions (CQs), covering six distinct domains. The dataset provides a gold-standard OWL ontology, two CQ files (corresponding to two evaluation tasks), and annotation tables used during construction for each domain. It includes two main tasks: CQ2Term (extracting all possible classes and properties from a given CQ) and CQ2Onto (generating a complete OWL ontology from a set of CQs). The dataset comprises six domain ontologies of varying sizes (Wine, AWO, ODRL, Water, VGO, and SWO), with each domains data including input files for both tasks, gold-standard ontology files (in OWL format), and Excel sheets documenting the annotation process. File formats include JSON for CQ lists and RDF/XML for OWL ontology files. The dataset is suitable for research and evaluation in ontology engineering, ontology generation, and term extraction tasks.




