Multimodal fake news datasets

We evaluate the proposed FSDUF model on three publicly available social media benchmark datasets: Weibo {jin2017multimodal}, Twitter {boididou2015verifying}, and Pheme {zubiaga2017exploiting}.Weibo dataset consists of original tweet texts and corresponding images collected from authoritative news sources in China.  Twitter dataset contains real and fake news samples sourced from the MediaEval Benchmarking Initiative. Pheme dataset includes rumors and non-rumors posted on Twitter during breaking news events, covering nine different events, with each rumor annotated as real or fake. All datasets are randomly divided into training, validation, and testing sets with a ratio of 7:1:2.