Enhancing Information Integrity: Machine Learning Methods for Fake News Detection

With growing and advancing technology, people have access to the Internet easily which leads to the availability of news online and can reach from one part of the world to another. Internet is the ideal environment for the growth and dissemination of malevolent and fake news. Fake news detection has become a serious problem in the recent era. Due to this, it has become important to detect whether a floating news is fake or real as it may have a serious negative impact on individuals and society. In this study, we use WELFake and Real and Fake News Classification Dataset from the Kaggle Repository for fake news classification using five machine learning algorithms: Naïve Bayes (NB) classifier, decision tree (DT), random forest (RF), support vector machine (SVM), and K-nearest neighbor (KNN). Five metrics are used: accuracy, precision, recall, F1-score, and AUC for comparing the model’s performances. We notice that the best-performing models are DT and RF yielding an accuracy of 93.92% and 91.18% respectively.