Fake News Detection Using Machine Learning Algorithm

Abstract

Fake news is a big concern since it spreads widely over social media and other media channels, creating significant social and national damage with devastating consequences. In order to address this issue, substantial research how to be recognized on fake news detection has been done. The purpose of this study is to analyze the existing research on fake news detection and select the best conventional machine learning frameworks to develop an algorithm for supervised machine learning. The algorithm will be able to classify false reports as true or fraudulent using textual analysis methods such as NLP. The proposed process involves data preprocessing and vectorization, where the NLP library will be used to perform tokenization and feature extraction of text data, utilizing tools such as Count Vectorizer and Tiff Vectorizer. Further, feature selection methods will be employed to evaluate and determine the most fitting features to achieve the highest precision based on confusion matrix results. The outcome of the decision tree classifier algorithm gives the accuracy of 90%.