DBPedia900K

该数据集是从DBPedia中提取的一个子集，包含了135万个三元组、881,649个实体和83个谓词。为了提高与用户查询的相关性，该数据集经过处理，特别关注基于用户查询日志的相关三元组。此外，它还根据用户查询日志进行了筛选，以增强数据的相关性。该数据集的规模达到了135万个三元组，其任务是知识图谱的完善。

This dataset is a curated subset extracted from DBPedia, containing 1.35 million triples, 881,649 entities and 83 predicates. To enhance its relevance to user queries, the dataset has been processed to specifically prioritize triples relevant to user query logs. Additionally, it has been filtered based on user query logs to further improve data relevance. With a total of 1.35 million triples, the downstream task targeted by this dataset is knowledge graph completion.