Fake News Detection on Social Media Data using Community Notes with Machine Learning

Abstract

The spread of misinformation on digital platforms has become a critical issue that requires the development of effective detection mechanisms. This study explores the use of machine learning techniques to classify fake and real news using data sourced from Platform X and the Disinformation Combat Center (DMM). The data set consists of approximately 7,000 fake and 17,000 real news samples, which are processed through data cleaning, labeling, and transformation techniques such as TFIDF vectorization. Various classification models, including Naive Bayes, Random Forest, Support Vector Machine (SVM), and Logistic Regression, are employed to evaluate the effectiveness of different approaches. The study further examines the impact of class balance on model performance, comparing results from balanced and imbalanced datasets. The findings contribute to ongoing research on misinformation detection by providing insight into the most effective methodologies for automated fake news classification.