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

Fake News Detection: Covid-19 Perspective Global Journal of Computer Science and Technology Volume XXII Issue II Version I 10 Year 2022 ( ) C © 2022 Global Journals Table 3: Fake News Detection Result Using Count Vectorizer presenting a data set containing more than 2900 fake and true news on Covid-19. In this digital era, people are mostly dependent on online news sources than printed news. As there are lots of news sources available online. So, there is a high risk of getting fake news on the online platform and fake news detection using the machine learning techniques can reduce the immense spread. To boost solving this issue, we have collected a new data set from traditional media based on Covid-19. We hope that This data set will help future researchers to contribute more to detecting fake news. We have also analyzed our data by four different existing supervised algorithms. We have calculated accuracy, recall, f1- score, and precision regarding count vectorizer and TF- IDF to detect fake news using algorithms and our data set on Covid-19. We have analyzed the algorithm with the four different amount of data. We have also analyzed how excessive amounts of one type of data can severely affect detection systems. As in this work, we only focused on the news related to Covid-19 and the news are from the traditional media so for this the number of data decreased. A largescale data set can be more effective in building a fake news detection system. As the fake news collection is challenging itself, especially from traditional media as there is a little amount of fake news in fact-checking sites. Also, we have faced similar kinds of fake news on different fact-checking sites. In the year 2020, many more events happened besides Covid-19. So, a positive path is to build a robust and large-scale data set of fake news from different media like social media, which researchers will use to facilitate additional study in this field. Using unsupervised or semi-supervised learning while detecting fake news can show better performance. R eferences R éférences R eferencias 1. Abdullah-All-Tanvir, Ehesas Mia Mahir, Saima Akhter, and Mohammad Rezwanul Huq. Detecting fake news using machine learning and deep learning algorithms. In 2019 7th International Conference on Smart Computing Communications (ICSCC), pages 1–5, 2019. 2. Vasu Agarwal, H. Parveen Sultana, Srijan Malhotra, and Amitrajit Sarkar. Analysis of classifiers for fake news detection. Procedia Computer Science, 165:377 – 383, 2019. 3. Sarah A. Alkhodair, Steven H.H. Ding, Benjamin C.M. Fung, and Junqiang Liu. Detecting breaking news rumors of emerging topics in social media. Technique Algorithms T/F Support Precision Recall F1-Score Accuracy Count Vectorizer Multinomial na¨ıve bias Set-1 0 53 0.89 0.77 0.83 0.85 1 63 0.83 0.92 0.87 Set-2 0 63 0.89 0.86 0.87 0.88 1 69 0.89 0.86 0.87 Set-3 0 61 0.88 0.80 0.84 0.88 1 104 0.89 0.93 0.91 Set-4 0 53 0.79 0.77 0.78 0.91 1 211 0.94 0.95 0.95 Passive Aggressive Set-1 0 53 0.89 0.89 0.89 0.90 1 63 0.90 0.90 0.90 Set-2 0 63 0.89 0.86 0.87 0.88 1 69 0.87 0.90 0.89 Set-3 0 61 0.84 0.84 0.84 0.88 1 104 0.90 0.90 0.90 Set-4 0 53 0.80 0.68 0.73 0.9 1 211 0.92 0.96 0.94 Support Vector Machine Set-1 0 53 0.82 0.85 0.83 0.84 1 63 0.87 0.84 0.85 Set-2 0 63 0.86 0.87 0.87 0.87 1 69 0.88 0.87 0.88 Set-3 0 61 0.84 0.84 0.84 0.88 1 104 0.90 0.90 0.90 Set-4 0 53 0.80 0.68 0.73 0.90 1 211 0.92 0.96 0.94 Logistic Regression Set-1 0 53 0.89 0.79 0.84 0.86 1 63 0.84 0.92 0.88 Set-2 0 63 0.90 0.86 0.88 0.89 1 69 0.88 0.91 0.89 Set-3 0 61 0.88 0.82 0.85 0.89 1 104 0.90 0.93 0.92 Set-4 0 53 0.92 0.62 0.74 0.91 1 211 0.91 0.99 0.95