WiKC

该数据集是对Wikidata分类法的精简版，采用大型语言模型和图挖掘技术自动清理，解决了诸如歧义、不准确、循环和冗余等问题。与YAGO 4.5分类法相比，WiKC在实体类型识别方面表现出显著的准确性提升，相较于原始的Wikidata分类法有了明显的改进。该任务涉及实体类型识别和分类法精炼。

This dataset is a streamlined version of the Wikidata taxonomy, which is automatically cleaned using large language models and graph mining techniques to address issues such as ambiguity, inaccuracy, cycles, and redundancy. Compared with the YAGO 4.5 taxonomy, WiKC exhibits significant accuracy improvements in entity typing, with notable enhancements over the original Wikidata taxonomy. This task involves entity typing and taxonomy refinement.