Table 8.
The features and datasets used in the hybrid approaches
| Feature and metadata | Datasets | Reference |
|---|---|---|
| Features and textual metadata of the news content: title, content, date, source, location | SOT fake news dataset, LIAR dataset and FA-KES dataset | Elhadad et al. (2019) |
| Spatiotemporal information (i.e., location, timestamps of user engagements), user’s Twitter profile, the user engagement to both fake and real news | FakeNewsNet, PolitiFact, GossipCop, Twitter | Nyow and Chua (2019) |
| The domains and reputations of the news publishers. The important terms of each news and their word embeddings and topics. Shares, reactions and comments | BuzzFeed | Xu et al. (2019) |
| Shares and propagation path of the tweeted content. A set of metrics comprising of created discussions such as the increase in authors, attention level, burstiness level, contribution sparseness, author interaction, author count and the average length of discussions | Twitter dataset | Aswani et al. (2017) |
| Features extracted from the evolution of news and features from the users involved in the news spreading: The news veracity, the credibility of news spreaders, and the frequency of exposure to the same piece of news | Twitter dataset | Previti et al. (2020) |
| Similar semantics and conflicting semantics between posts and comments | RumourEval, PHEME | Wu and Rao (2020) |
| Information from the publisher, including semantic and emotional information in news content. Semantic and emotional information from users. The resultant latent representations from news content and user comments | Guo et al. (2019) | |
| Relationships between news articles, creators and subjects | PolitiFact | Zhang et al. (2020) |
| Source domains of the news article, author names | George McIntire fake news dataset | Deepak and Chitturi (2020) |
| The news content, social context and spatiotemporal information. Synthetic user engagements generated from historical temporal user engagement patterns | FakeNewsNet | Shu et al. (2018a) |
| The news content, social reactions, statements, the content and language of posts, the sharing and dissemination among users, content similarity, stance, sentiment score, headline, named entity, news sharing, credibility history, tweet comments | SHPT, PolitiFact | Wang et al. (2019a) |
| The source of the news, its headline, its author, its publication time, the adherence of a news source to a particular party, likes, shares, replies, followers-followees and their activities | NELA-GT-2019, Fakeddit | Raza and Ding (2022) |