Snopes and Politifact Datasets

该数据集包含了从两个主要的事实核查网站——Snopes和Politifact——收集的文本主张及其可信度标签的元组。每个主张都根据其来源被标注为真实或虚假。其中，Snopes网站上的主张被直接标注为真实或虚假，而Politifact网站上的主张原本有六个标签，为了这个数据集的目的，这些标签被合并为真实或虚假。该数据集的任务是进行虚假新闻检测。

This dataset comprises tuples of textual claims and their corresponding credibility labels, collected from two prominent fact-checking websites: Snopes and Politifact. Each claim is labeled as either true or false based on its source. Specifically, claims from Snopes are directly labeled as true or false, while the original Politifact claims have six distinct labels, which are consolidated into two categories: true and false, for the purpose of this dataset. The task of this dataset is fake news detection.

搜集汇总

背景与挑战

背景概述

该数据集包含Snopes和PolitiFact两个子集，用于证据感知的假新闻检测研究，支持训练和评估分层多头注意力网络模型。数据集与假新闻验证相关，提供了格式化数据用于机器学习实验，旨在提高假新闻识别的准确性和可解释性。