five

Sentiment Text Classification and Part of Speech (SCPOS) datasets

收藏
arXiv2023-09-14 更新2024-08-06 收录
下载链接:
http://arxiv.org/abs/2309.03787v2
下载链接
链接失效反馈
官方服务:
资源简介:
SCPOS数据集是由横滨国立大学的研究团队创建,旨在支持情感分析研究,特别是探索文本情感极性和词性情感极性之间的关系。该数据集包含四个子集,每个子集都基于现有的情感分类数据集进行标注,重点关注影响文本整体情感极性的词汇。数据集的创建过程中,研究团队采用了手动标注和基于规则的方法,确保数据的高质量和准确性。SCPOS数据集的应用领域主要集中在情感分析和词性标注,旨在通过分析词汇的情感极性来提高情感分类模型的性能。

The SCPOS dataset was developed by a research team from Yokohama National University to support sentiment analysis research, particularly the exploration of the relationship between text-level sentiment polarity and part-of-speech (POS) based sentiment polarity. This dataset comprises four subsets, each annotated based on existing sentiment classification datasets, with a core focus on lexical items that affect the overall sentiment polarity of the text. During the dataset construction, the research team employed both manual annotation and rule-based approaches to guarantee the high quality and accuracy of the data. The primary application domains of the SCPOS dataset are sentiment analysis and part-of-speech tagging, with the objective of enhancing the performance of sentiment classification models by analyzing the sentiment polarity of vocabulary.
提供机构:
横滨国立大学, 日本
创建时间:
2023-09-07
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作