Fake news detection on social networks using Machine learning techniques

Abstract

The constant circulation of fake news directly or indirectly produces a huge negative impact on vast majority of the society. Users throughout the world are actively visiting well-known social networking sites such as Facebook, Twitter, Instagram, LinkedIn, and others. The majority of online social networks authentication systems are insufficient and rely on only general information such as the user's name, photo, and location. Because of the system's flaws, malevolent users can simply abuse users' information, cloning their identities and disseminating false information. The major thought employed before implementing this project was to extend the scope of application of machine learning techniques and detection of fake news model on the English language is proposed by the use of machine learning techniques. It is investigated and compared with three different evaluation models namely Count Vectorizer, TF-IDF Vectorizer and N-gram and four machine learning techniques were used namely Naive Bayes, SVM, Random Forest, Logistic Regression. The proposed strategy achieves the maximum accuracy % when using TF-IDF highlights and an SVM classifier. The highest level of precision is 93%.