factnet_relations
收藏资源简介:
FactNet关系数据集包含FactSynset之间丰富的语义关系,支持高级推理和跨语言事实检索。这些关系涵盖了上下位关系、因果关系、时间关系、地理关系等多种语义连接。数据集采用parquet格式存储,主要字段包括关系ID、源/目标FactSynset ID、关系类型、置信度、支持该关系的事实陈述ID、检测方法以及元数据。关系类型多样,包括等价关系、上下位关系、因果关系、空间关系、部分关系、成员关系、时间序列关系、影响关系等20余种。该数据集支持多跳事实推理、因果与时间推断、地理空间推理、语义相似度计算等高级应用场景。数据来源于Wikidata和Wikipedia,采用CC BY-SA许可协议。
FactNet relational dataset contains rich semantic relationships between FactSynsets, supporting advanced reasoning and cross-lingual factual retrieval. These relationships cover various semantic connections including hyponymy, hypernymy, causality, temporal relations, geographic relations and more. The dataset is stored in Parquet format, with its core fields including relationship ID, source and target FactSynset IDs, relationship type, confidence score, factual statement IDs supporting the relationship, detection method, and metadata. There are over 20 types of relationships, such as equivalence, hyponymy, hypernymy, causality, spatial relations, part-whole relations, membership relations, temporal sequence relations, influence relations and others. This dataset supports advanced application scenarios including multi-hop factual reasoning, causal and temporal inference, geospatial reasoning, semantic similarity calculation and more. The data is sourced from Wikidata and Wikipedia, and is distributed under the CC BY-SA license.




