Global Journal of Computer Science and Technology, C: Software & Data Engineering, Volume 22 Issue 2

Fake News Detection: Covid-19 Perspective Abstract- The development of social media has contributed to a remarkable rise in the spread of fake news. Today people rely more on online news outlets. The chance of receiving fake news on an online platform is high. As we went through a pandemic and the Covid-19 was the most absorbing topic of 2020, much news on Covid-19 was published every day in traditional media and social media. Among that news, some are fake. In this work, we have collected a new dataset for detecting fake news from traditional media on Covid-19. We have gathered more than 3000 pieces of news from traditional media out of the 170 are fake ones that were collected from fact-checking sites. Then we have tested the existing four classification algorithms with our dataset using Count Vectorizer and TF- IDF. We have merged 170 fake news with four scales of true news and analyzed the outcome. fake news, fake news detection, traditional media, data set, covid-19, social media. I. I ntroduction n 2020, we have gone through trauma due to covid- 19. This pandemic is not less than a trauma. Covid- 19 has become a challenge to each and everyone whose country is facing a huge number of positive cases regarding this and the number of deaths are rising on a daily basis, which actually makes one depressed. During this pandemic situation people are stuck in their own places thus they are consuming much more internet than before. According to openvault, data usage has increased 47 percent during this pandemic [34]. As people are consuming more internet, they are also getting their news from it. Among those news stories, some of them are fake, but it is hard for maximum people to find out whether the news is fake or not. Through an automated system we can easily detect fake news. An automated system could be defined as a technique that performs a task using programming inputs and computerized feedback control to verify that the instructions are followed correctly. The end result is a technology that can function without any need for human interaction [18]. An automated system needs a well-labeled data set to detect fake news from tons of news. So, we have made a new data set on Covid-19. What is Fake News? Fake news is misleading or false news information introduced as news [19]. There is no valid definition of fake news. Many researchers [32], [11] have adopted Hunt Allcot et al. definition. They defined, a news article that is deliberately and verifiably false and could deceive readers is called fake news [4]. Also another explanation from S. Desai, fake news is those articles that are fabricated, false information with no supportable facts, origin, or quotes [13]. These two definitions have two key features: authenticity, fake news includes false information, facts, sources, or quotes to make it more authentic so that readers can have no doubt about the news and another is intent. Basically, fake news is generated with the purpose of deceiving readers, sometimes to harm individuals. In some papers, satire or parody news are also considered as fake news [7], [8]. Fake news can be spread via two types of media such as traditional media and social media [32]. Traditional media are newspapers, magazines, blogs, etc. and social media are Facebook, Twitter, WhatsApp, etc. So basically, we have focused our data set to detect fake news from traditional media and news related to Covid-19. According to researchers, in the first three months of the year 2020 at least 800 people might have died across the world because of fake news about Covid-19 [10]. Fake news is not only spreading disinformation or rumors, but also taking valuable lives and also hampering public properties. To check the facts of fake news there are some delegated websites like Politifact [20], lead stories [15], FactCheck [28], the full fact [24], health feedback [35] where they manually updates potential fake news stories which are published online platforms. We have collected a vast number of news data sets so that our outcome can help the upcoming researchers to work on this problem. We have worked on Covid-19 which is the most hooked and challenging topic in the world and every day we have got thousands of news regarding this pandemic. From those news some of them are fake. In the beginning, people used to believe in every news regarding Covid- 19. By these, some of them became so panic that they were about to lose their lives and some of them already lost theirs. There were too many rumors that people started believing those. For example, in Iran, there were rumors across the country that alcohol consumption would help combat Covid-19 [16], [5]. I Global Journal of Computer Science and Technology Volume XXII Issue II Version I 1 Year 2022 ( ) C © 2022 Global Journals Index Terms: Md. Ziaur Rahman Shamim α , Shaheena Sultana σ , Anika Tabassum ρ , Israt Tabassum Ѡ & Sarkar Binoyee Farha ¥ Author α σ ρ Ѡ ¥: Department of Computer Science and Engineering, Notre Dame University Bangladesh Dhaka, Bangladesh. e-mails: zrshamim8822@gmail.com , shaheenacse@ndub.edu.bd, anika4725@gmail.com , israt.tabassum34@gmail.com, farha1binoyee@gmail.com