turkish-fake-news-detection

MiDe22仓库专注于土耳其语推文的多事件标注，用于错误信息检测。数据内容为5064条推文，涵盖2020年至2022年间的多个事件，包括俄乌战争、COVID-19疫情和难民问题。推文被标注为“真”、“假”和“其他”三类，以区分正确信息、错误信息和无法归类的推文。数据采集后经过五位标注员的标注，并通过Krippendorf’s alpha系数评估标注一致性。该仓库采用MIT许可。

The MiDe22 Repository focuses on multi-event annotated Turkish tweets for misinformation detection tasks. The dataset contains 5,064 tweets covering multiple events spanning from 2020 to 2022, including the Russia-Ukraine War, the COVID-19 pandemic, and the refugee issue. Each tweet is categorized into three classes: "True", "False", and "Other", to distinguish genuine information, misinformation, and unclassifiable tweets. Following data collection, the tweets were annotated by five annotators, and the annotation consistency was evaluated using Krippendorf's alpha coefficient. This repository is licensed under the MIT License.