franciellevargas/FactNews

FactNews是一个葡萄牙语的大型句子级标注语料库，用于真实性预测。该数据集包含6191个句子，根据AllSides提出的真实性定义和媒体偏见定义进行标注。FactNews用于评估新闻来源的整体可靠性，通过制定两个文本分类任务来预测新闻报道的句子级真实性和媒体机构的偏见。实验表明，带有偏见的句子比真实句子包含更多的词汇，并且情感表达更为突出。对新闻文章主观性和公正性的细致分析显示出预测整个媒体机构可靠性的前景。

FactNews is a large sentence-level annotated corpus for factuality prediction in Portuguese. The dataset consists of 6,191 sentences annotated according to the factuality and media bias definitions proposed by AllSides. FactNews is used to assess the overall reliability of news sources by formulating two text classification tasks for predicting the sentence-level factuality of news reporting and the bias of media outlets. Experiments show that sentences with bias contain a higher number of words compared to factual sentences, and there is a predominance of emotions. The fine-grained analysis of subjectivity and impartiality in news articles has shown promising results for predicting the reliability of entire media outlets.