XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE Machine Learning Approaches for Detecting and Mitigating the Impact of Fake News in Online Information Ecosystems Temitayo Elijah Balogun Department of Information Systems Federal University of Technology, Akure Akure, Nigeria 0000-0003-0622-8965 Oluwaseyifunmi Balogun Department of Computer Science Obafemi Awolowo University, Ile Ife Ile-Ife, Nigeria 9 Samson Adebowale Abosede Department of Computer Science Federal College of Agriculture, Akure Akure, Nigeria adebowale4st@gmail.com Paul Kehinde Olotu Department of Information Technology Federal University of Technology, Akure Akure, Nigeria pkolotu@futa.edu.ng Shade Christiana Joda Department of Computer Science Federal College of Agriculture, Akure Akure, Nigeria shadejoda@gmail.com Samuel Gbenga Faluyi Department of Computer Science Ekiti State Polytechnic, Isan Ekiti, Nigeria https://orcid.org/0000-0002-2662-1599 Abstract—This research investigates the pervasive issue of fake news, particularly its influence on societal behavior and the historical roots dating back to the advent of the printing machine in 1439. Focusing on the primary goal of influencing public opinion, the study aims to address the challenges posed by misinformation in the digital age, where approximately 62% of US citizens obtain news from social media. The study proposes leveraging Support Vector Machine and Random Forest for the detection and categorization of false news articles, considering two key approaches—statistical analysis and knowledge base searching. By training the model on examples of both true and fake news using a dataset from Kaggle with 4009 instances of real and fake news, this research seeks to assess the accuracy of the models’ predictions. From the research, the Support Vector Machine Model achieved the best accuracy with 98%. The outcomes aimed to contribute to the development of effective tools for combatting the proliferation of fake news, particularly highlighting the role of machine learning in enhancing detection methods. Keywords—Machine Learning, Fake News Detection, Support Vector Machine, Random Forest, Online News. I. INTRODUCTION Social media's increasing dominance in news consumption presents a double-edged sword, offering fast access but exposing users to a constant influx of fake news. Identifying and debunking misinformation is crucial at both individual and societal levels, necessitating robust fake news detection systems. Challenges arise from the deceptive construction of fake news, aiming to shape public opinion based on social and political situations [1]. The proliferation of misinformation, deliberately disseminated to manipulate public behavior, has emerged as a societal challenge. The impact of fake news on significant events, exemplified by its influence on the 2016 US presidential elections and the "Brexit" referendum, underscores its pervasive reach [2]. With approximately 62% of US citizens relying on social media for news [3], the potency of fake news is further amplified, as evidenced by its increased shares on platforms like Facebook [4]. This historical root of misinformation trace back to the introduction of the printing machine in 1439, marking fake news as a longstanding issue [5]. Defined by authenticity and intent, fake news embodies erroneous information intentionally crafted to deceive [6]. Noteworthy is its historical application in influencing public opinion, as seen in the one-sided party newspapers of the 19th century. Recognizing the diverse motives behind fabricating news—such as influencing public opinion, generating revenue through clickbait, and satirical writing—this research narrows its focus to the primary goal of ensuring the populace can distinguish between fake and real news [7], [8]. Real-world consequences, like the attack on a pizzeria resulting from false news circulation [9], emphasize the urgency of addressing the harmful repercussions. Machine learning has played a key role in identification, classification and detection across diverse fields in recent times spanning from health [10], [11] to movies [12] to building structures [13]. Ascertaining whether machines can surpass human capabilities in tackling the fake news problem is a pivotal question [14]. Figure 1: various approaches to fake news detection [6]. The machine learning models employed in this research, aim to identify online news likely to be false reports. Through machine learning methods, this study seeks to evaluate the efficacy and limitations of these models in detecting fake news patterns, excluding external knowledge. This research will compare machine learning models, including Support Vector Machine, and Random Forest Classifiers, to identify the most effective approach. 2024 International Conference on Science, Engineering and Business for Driving Sustainable Development Goals (SEB4SDG) | 979-8-3503-5815-5/24/$31.00 ©2024 IEEE | DOI: 10.1109/SEB4SDG60871.2024.10630248 Authorized licensed use limited to: Durban University of Technology. Downloaded on August 27,2024 at 17:06:26 UTC from IEEE Xplore. Restrictions apply.