ASOIF和HP系列小说词汇嵌入评估数据集
收藏arXiv2019-03-07 更新2024-06-21 收录
下载链接：
https://github.com/gwohlgen/digitalhumanities_dataset_and_eval
下载链接
链接失效反馈官方服务：
资源简介：
本数据集由俄罗斯圣彼得堡国立信息技术机械与光学大学软件工程与计算机系统学院的研究团队创建，专注于评估在数字人文领域中词汇嵌入技术的性能。数据集基于两部流行的奇幻小说系列：乔治·R·R·马丁的《冰与火之歌》和J.K.罗琳的《哈利·波特》。数据集包含31474个评估单元，用于测试词汇相似性和词汇侵入等任务。创建过程中，研究团队使用了多种词汇嵌入模型，并手动构建了高质量的测试数据集。该数据集旨在为未来的研究提供基准，并解决特定领域和较小语料库中词汇嵌入模型的性能问题。
This dataset was created by a research team from the School of Software Engineering and Computer Systems, National Research University of Information Technologies, Mechanics and Optics (ITMO University), St. Petersburg, Russia, and focuses on evaluating the performance of word embedding techniques in the digital humanities domain. The dataset is built upon two popular fantasy novel series: *A Song of Ice and Fire* by George R. R. Martin and *Harry Potter* by J. K. Rowling. It contains 31,474 evaluation units for tasks such as word similarity and word intrusion. During the dataset construction process, the research team adopted multiple word embedding models and manually built a high-quality test dataset. This dataset aims to provide a benchmark for future research and address the performance issues of word embedding models in specific domains and small-scale corpora.
提供机构：
俄罗斯圣彼得堡国立信息技术机械与光学大学软件工程与计算机系统学院
创建时间：
2019-03-07
