RuDSI
收藏arXiv2022-09-28 更新2024-06-21 收录
下载链接:
https://github.com/kategavrishina/RuDSI
下载链接
链接失效反馈官方服务:
资源简介:
RuDSI是一个针对俄语的词义诱导(WSI)新基准数据集,由俄罗斯高等经济学院创建。该数据集通过手动注释和半自动聚类Word Usage Graphs(WUGs)生成,完全基于俄罗斯国家语料库的文本,不依赖外部词义。数据集包含24个目标词,每个词有35个句子样本,通过随机抽样从俄罗斯国家语料库中获取。RuDSI旨在测试非分级词义诱导系统,适用于研究计算语言学中的词义诱导问题,特别是对于俄语的处理。
RuDSI is a novel benchmark dataset for Russian Word Sense Induction (WSI), developed by the Higher School of Economics of the Russian Federation. It is generated through manual annotation and semi-automated clustering of Word Usage Graphs (WUGs), and is fully based on texts from the Russian National Corpus, without relying on external word senses. The dataset includes 24 target words, each with 35 sentence samples randomly sampled from the Russian National Corpus. RuDSI aims to evaluate non-hierarchical word sense induction systems, and is suitable for researching word sense induction problems in computational linguistics, particularly for Russian language processing.
提供机构:
俄罗斯高等经济学院
创建时间:
2022-09-28



