Fake News Detection Using Various Machine Learning Approaches

Abstract

Fake news is spreading quickly on social media and other platforms, which is quite concerning since it may have serious negative effects on both the national and societal levels. Extensive research efforts are currently underway to detect and combat this issue. This study surveys the existing research on fake news identification and studies the efficacy of traditional ML (Machine Learning) models to develop a supervised ML method that can accurately identify fake news as either true or false. To achieve this, tools such as NLP and Python sci-kit-learn for textual analysis will be utilized. The process will involve feature extraction and vectorization, with the proposal of employing the Python sci-kit-learn library for tasks like feature extraction and tokenization of text data. This library provides valuable devices including Count Vectorizer and Tiff Vectorizer. Additionally, feature selection methods will be employed to test and detect the most suitable features that yield the greatest precision, as examined by the confusion matrix results.