People in this day and age like to keep up with the most recent news on social media since it is economical, simple, and can be obtained in a short amount of time. However, social media can be a double-edged sword, since it may propagate fake news or information that cannot be relied upon. The dissemination of false information could adversely affect both individuals and society as a whole. Therefore, researchers gave their all to devise a system that could identify fake news before its publication. This article gives an overview of the most major attempts that have been made to construct a system that can filter online news and detect false news with a decent amount of accuracy. Data mining approaches were reviewed in the detection phase, including feature extraction and model construction.
TopIntroduction
Define the problem broadly, discuss it in depth, and add other people's perspectives (literature review) to the conversation to prove, disprove, or illustrate your position. social media is increasingly influencing our daily lives, and people often get their news from social media more than traditional news outlets. Several factors contribute to this change in consumption behaviors on social media platforms: For example, different internet platforms often provide more well-timed and cheaper news than traditional media. Additionally, it is simpler to share, comment on, and discuss the information with groups and other people on the internet (Shu et al., 2017). According to a study (Wakefield, J., 2016), social media outlets outperform television as the most important news source. Despite the advantages of social media, the value of news on the internet is lesser than in traditional media. However, it is cheaper and faster to offer news to distribute, which leads to large volumes of fake news. By the end of the presidential election, over 1 million tweets referred to the fake news “Pizzagate” (Wikipedia contributors, 2022). In 2016, the Macquarie dictionary named “Fake news” the word of the year based on the prevalence of this new phenomenon.
Fake news spreads widely and can seriously affect individuals and society. Firstly, fake news damages the credibility of the news ecosystem. For example, the most popular fake news during the 2016 U.S. presidential election spread more widely on Facebook than the most popular authentic mainstream news (Buzzfeed, 2022), (Figure 1). Secondly, fake news intentionally misleads consumers. Usually, misinformation content is created by publicists to increase governmental influence or convey political messages. According to some reports, Russia has created fake accounts and social bots to spread misleading news (Time, 2022). Thirdly, misinformation content changes how people interpret and react to real news. In some cases, fake contents is intended to make people distrustful and confused, hindering their ability to detect what is real from what is not (Nytimes, 2022).
Figure 1. Example of Fake News in U.S Presidential Election (Higgins et al., 2016)
To help mitigate the negative effects of fake news to gain the general public and the information ecosystem. Many studies tried to increase methods to automatically discover fake information on social media. Detecting fake information on social media poses several challenges. However, fake information itself isn’t always a brand-new problem. The rise of net-generated information on social media makes fake information a more powerful force that challenges traditional journalistic norms. Numerous traits of this hassle make it uniquely challenging for automatic detection. This problem has numerous characteristics that make it uniquely difficult for automatic detection. First, fake information is intentionally written to misinform readers, making it challenging to discover based on information content. Second, exploiting this auxiliary information results in another critical challenge, i.e., the quality of the information itself.
Fake news is generally associated with newly emerging, time-essential events, which might not have been proven effectively through current information stands due to the lack of corroborating indication or claims. Powerful strategies to differentiate credible users, extract beneficial post features, and take advantage of network interactions are an open region of research and need additional investigations.