Detecting Fake News Using Machine Learning Based Approaches

The spread of false information, commonly known as “fake news,” has become a significant problem in recent years, with the potential to mislead the public and influence important decisions. In this research, we focus on creating an automated system for fake news detection using natural language processing of news texts. We investigate different machine learning-based classification techniques to predict whether a text is a real or fake news. We utilized popular datasets from Kaggle and implemented Logistic Regression, Support Vector Machine, decision tree, k-Nearest Neighbors, multinomial naïve Bayes, and Multilayer Perceptron, as well as an ensemble technique called stacking, which utilizes the other models in its prediction. We also perform a comparative analysis of the accuracies of the different techniques in fake news detection. From our experimental analysis, we found that the ensemble learner, Support Vector Machine, and Multilayer Perceptron outperform the other approaches and have the highest overall accuracies.