FinnSentiment

FinnSentiment数据集由赫尔辛基大学创建，包含27000条从芬兰社交媒体网站Suomi24随机选取的句子。这些句子经过两位自动标注者和三位人工标注者的情感极性标注，旨在提供一个大规模的芬兰语情感分析资源。数据集的创建过程包括随机选择句子、使用自动标注工具进行初步筛选，以及人工标注者的独立标注。该数据集主要用于情感分析研究，特别是在社交媒体文本中识别和分析情感极性，以及评估不同标注方法的一致性和有效性。

The FinnSentiment dataset, developed by the University of Helsinki, contains 27,000 sentences randomly sampled from the Finnish social media platform Suomi24. All sentences were annotated for sentiment polarity by two automatic annotators and three human annotators, with the objective of creating a large-scale Finnish-language resource for sentiment analysis. The dataset construction process involves three core steps: random sentence selection, preliminary filtering via automatic annotation tools, and independent annotation carried out by human annotators. This dataset is primarily applied to sentiment analysis research, specifically for identifying and analyzing sentiment polarity in social media texts, as well as evaluating the consistency and effectiveness of various annotation methods.