Development of a Fake News Detection Model Using Decision Tree Algorithm

Abstract

Over 500,000 Americans are influenced each year by fake news, which has an impact on society. Researchers are examining data on fake news identification and looking into machine learning models for a quicker and more accurate categorization as part of their investigation of false news detection techniques. This study's main objective is to examine, contrast, and evaluate how well three distinct machine learning algorithms do at spotting fake news. Decision tree, Random forest, and logistic regression are the machine learning algorithms. There are 21418 datasets available for real news, compared to 23503 for fake news. The open-source Kaggle dataset, which was selected for the project, served as the source of both datasets. The results of the construction and comparison showed that the decision tree model did the best forecasting with a maximum accuracy of 99.64% and the random forest model performed well with an accuracy of 98.89%. However, the logistic regression model did well in this assignment with an accuracy of 98.88%. Using these models has the critical goal of improving the news verification system's accuracy and dependability. Keyword: Fake news, Verification System, Machine Learning Model. Proceedings Citation Format Adeola, O.O., Bello, G.R. Oluwasola, B.S. & Lateef, K.R. (2023): Development of a Fake News Detection Model Using Decision Tree. Proceedings of the Cyber Secure Nigeria Conference. Nigerian Army Resource Centre (NARC) Abuja, Nigeria. 11-12th July, 2023. Pp 81-88. https://www.csean.org.ng/. dx.doi.org/10.22624/AIMS/CSEAN-SMART2023P10