Combating the menace: A survey on characterization and detection of fake news from a data science perspective

• Published in: International journal of information management data insights Vol. 1; no. 2; p. 100052
• Main Authors: Ansar, Wazib, Goswami, Saptarsi
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.11.2021; Elsevier
• Subjects: COVID-19; Deep learning; Fake news; Fake news data sets; Machine learning; Transformer NLP; Deep learning; COVID-19; Fake news; Transformer NLP; Fake news data sets; Machine learning
• ISSN: 2667-0968; 2667-0968
• DOI: 10.1016/j.jjimei.2021.100052

•An interpretation of fake news has been presented supported by a comparison of its categories based upon aspects like objective, intention, distinguishability and information type.•A survey of related works has been conducted on the characterization, feature extraction and subsequent detection of fake news using statistical, machine learning and deep learning approaches.•An analysis of the renowned fake news detection repositories has been performed based upon various parameters.•A case study on tweets related to COVID-19 and distinguishing between factual and emotional tweets. Factors behind the proliferation of fake news during the pandemic and the efforts to impede this spread of misinformation with associated research works. Journalism has always remained a vital constituent of our society and journalists play a key role in making people aware of the happenings and developments in society. This spread of information enables shaping the ideologies, orientations and thoughts of individuals as well as the society. Contrary to this, the spread of misinformation or fake news leads to detrimental consequences. With the advent of social media, the menace of fake news has become grievous due to the unrestrained propagation of information and difficulty to track several accounts operated by humans or bots. This menace can be mitigated through data science approaches by combining artificial intelligence with statistics and domain-based knowledge. In this paper, a survey of works aimed at characterization, feature extraction and subsequent detection of fake news has been conducted from a data science perspective. Along with it, an analysis of the 8 renowned fake news detection repositories has been presented. Furthermore, through a case study on tweets related to COVID-19 pandemic, the factors behind the spread of misinformation during critical times, distinguishing between factual and emotional tweets and viable approaches to restrain fake news has been enunciated.