UPFD

该数据集名为UPFD，包含了两组来自Twitter的树状结构假新闻和真新闻传播图，专注于用户偏好意识的假新闻检测。数据集包括四种不同的节点特征类型：个人资料、Spacy、Bert和内容。规模上，这两组图分别代表了假新闻和真新闻的传播情况，任务旨在进行用户偏好意识的假新闻检测（Upfd）。

The dataset is named UPFD, which contains two sets of tree-structured fake and real news propagation graphs sourced from Twitter, focusing on user-preference-aware fake news detection. It includes four distinct types of node features: profile, Spacy, Bert, and content. In terms of scale, the two graph sets respectively represent the propagation scenarios of fake news and real news. The task targets user-preference-aware fake news detection (Upfd).

搜集汇总

背景与挑战

背景概述

UPFD数据集是一个用于假新闻检测的图结构数据集，包含来自Politifact和Gossipcop的新闻传播网络，支持多种节点特征类型。数据集通过图分类任务形式呈现，适用于图神经网络模型的训练和评估。