Fake News Detection Model Basing on Machine Learning Algorithms

Abstract

The rapid growth of the internet and easy communication has made it quick and simple to create and spread news. Social media users now generate and share more information than before, but some of it is false and unrelated to reality. Detecting false information in text is challenging, even for experts who need to consider multiple factors to determine authenticity. Malicious misinformation on social media negatively affects societies, especially during crises like terrorist attacks, riots, and natural disasters. To minimize the harmful impact, it is crucial to identify rumors quickly. This study aims to build a learning model for detecting fake news. This research paper relies on finding and analyzing the characteristics of the text, then the words are converted into features using TF-IDF technology, after that the highest-ranking features are identified for the purpose of studying and distinguishing the spread of news, whether it is real or fake using machine learning techniques. Finally, the Logistic Regression, Decision Tree, Gradient Boosting and Random Forest algorithm has been adapted. The accuracy of Logistic Regression is 0.985, Random Forest (0.989) whereas the accuracy of Decision Tree is 0.994 and Gradient Boosting (0.9949), respectively.