COVID-19 Fake News Dataset (COVID19 Fake News Detection in English)

除了新型冠状病毒肺炎大流行，我们还在与 “信息demic” 作斗争。假新闻和谣言在社交媒体上猖ramp。相信谣言会造成重大伤害。在大流行之时，这种情况进一步加剧。为了解决这个问题，我们策划并发布了一个手动注释的数据集，其中包含10,700关于COVID-19的社交媒体帖子和真实和虚假新闻的文章。我们使用四个机器学习基线 (决策树，逻辑回归，梯度提升和支持向量机 (SVM)) 对带注释的数据集进行基准测试。用SVM F1-score，我们获得了93.46 \ % 的最佳性能。In addition to the COVID-19 pandemic, we are also grappling with an 'infodemic'. Fake news and rumors run rampant on social media platforms, and believing such rumors can cause significant harm. This situation has been further exacerbated amid the COVID-19 pandemic. To address this issue, we curated and released a manually annotated dataset containing 10,700 entries of COVID-19-related social media posts and articles of real and fake news. We benchmarked the annotated dataset using four machine learning baselines: decision trees, logistic regression, gradient boosting machines, and support vector machines (SVM). We achieved the optimal performance of 93.46% using the F1-score metric for the SVM model.

搜集汇总

背景与挑战

背景概述

该数据集专注于COVID-19假新闻检测，旨在提供英语文本数据以支持相关研究。它是一个公开的小型数据集（1.3KB），由OpenDataLab发布，适用于自然语言处理和虚假信息分析领域。