Performance Analysis of ML Techniques in Identification of Fake News Reshma Vunnava 1 , Lakshmikanth Bodla 2 , Mohan Kumar Dehury 3 , Bhabendu Kumar Mohanta 4 1,2,3,4 Department of CSE, Koneru Lakshmaiah Education Foundation, Vaddeswaram, Guntur - 522502, Andhra Pradesh, India Email: rreshu777@gmail.com 1 , lakshmikanthhbodla@gmail.com 2 , mohankdehury@gmail.com 3 , bhabendukumar@gmail.com 4 . Abstract—Most people are choosing to get their news via the internet since it is convenient and inexpensive, yet this leads to a rapid spread of fake news. Data is extremely vital in today’s world, and by 2023, 120 zeta bytes of data will be released per second. Many technologies are changing the world as a result of this massive volume of data. As the Internet has become increasingly popular, people rely on online news sources to keep up with the latest developments. With the development of the usage of platforms for social media such as Instagram, Facebook and Wikipedia, the news spread rapidly to users around the world in a short period of time. This may also lead to spread of fake news that can affect the society and individuals. In this paper, we have used Machine learning (ML) in detection of fake news. With the aid of ML techniques, we seek to conduct binary categorization of various news items available online in this work. Also, we proposed a fake news detection architecture and using that we presented a comparison of different ML techniques for fake news detection. Index Terms—Machine Learning, Classification, Fake News, Internet, Social Media. I. I NTRODUCTION ”Fake news” is a word that refers to propaganda that includes misinformation and is distributed through conventional and unconventional media outlets such as print and television, as well as social media [1]. The dissemination of such information is frequently motivated by a desire to mislead readers, harm an organization’s image, or profit for dramatic news and it is more often seen as one of the most serious challenges for democracy, freedom of speech, and the Western system. The world is evolving at a tremendous speed. There are numerous benefits for the living beings in this digital era, but there are also drawbacks. In today’s digital environment, there are several obstacles. Fake news is one of them. Fake news is very easy to propagate. False information is being circulated in order to smear the reputation of individuals or organizations [2]. It might be a piece of propaganda for a political party or group. Fake news may be spread through a variety of internet sites. This includes social media sites like Facebook and Twitter. Our perspective in the world is formed by the data we have processed. There is expanding proof that customers have reacted absurdly to messages that later ended up being phony. The current case is another kind of Covid plague, with counterfeit reports on the beginning, type, and conduct of the infection being conveyed over the Web. The circumstance deteriorated as more individuals read with regards to counterfeit substance on the web. Recognizing such messages online can be an overwhelming undertaking. Luckily, there are a few number juggling strategies that can be utilized to stamp a specific article as fake because of the substance of the text. The greater part of these procedures uses reality checking sites like Politi Truth and Snopes. There are a few stories kept by analysts that include a rundown of sites that have been recognized as obscure and bogus. One disadvantage of these checking methods is that they require human mastery to recognize the thing or site as fake. Furthermore, the reality checking site contains articles in explicit areas such as legislative issues. So, distinguishing fake news or stories from different areas, like diversion, sports, and innovation becomes difficult. Also, the Internet contains information in different arrangements like archives, videos, and sounds. Online news posted in unstructured arrangements (news, Proceedings of the International Conference on Sustainable Computing and Data Communication Systems (ICSCDS-2022) IEEE Xplore Part Number: CFP22AZ5-ART; ISBN: 978-1-6654-7884-7 978-1-6654-7884-7/22/$31.00 ©2022 IEEE 276 2022 International Conference on Sustainable Computing and Data Communication Systems (ICSCDS) | 978-1-6654-7884-7/22/$31.00 ©2022 IEEE | DOI: 10.1109/ICSCDS53736.2022.9760905 Authorized licensed use limited to: International Institute of Information Technology Bhubaneswar. Downloaded on May 02,2022 at 07:32:33 UTC from IEEE Xplore. Restrictions apply.