DBpedia English 2015 Dataset

该数据集包含了来自DBpedia的实体，并根据它们的类型进行分类，这有助于评估概念嵌入的质量。实体是基于它们的rdf:type被检索的，并且使用了诸如skip-gram、CBOW和GloVe等不同的模型来计算嵌入。该数据集的规模包括12个概念，每个概念有10,000个实例。其任务是评估概念嵌入的分类方面表现。

This dataset contains entities sourced from DBpedia, which are categorized by their types to facilitate the evaluation of concept embedding quality. The entities are retrieved based on their rdf:type, and various models such as skip-gram, CBOW, and GloVe are employed to compute the embeddings. The dataset consists of 12 concepts, with 10,000 instances per concept. The task is to evaluate the classification performance of concept embeddings.