Abstract:
Today, fake news detection is a trendy topic that has drawn a lot of attention from academics around the globe. Any content that is untrue and created with the intention of leading readers to believe a lie is typically considered fake news.
In this paper, a framework is proposed that should be used to begin the project. It calls for the application of text-processing, cleaning, and feature-extraction techniques to reorganize the information, which should then be "obeyed" into each classification model during training and parameter tuning to produce the most accurate and optimized predictions for identifying fake news.
This study examines three example datasets to better understand the background for identifying fake news. It also makes an effort to determine linguistic differences between false and real news items using a variety of visualization techniques. This text's goal is to provide a detailed analysis of the results of several popular machine learning classifiers, including the Support Vector Machine, the Naive Bayes Method, the Decision Tree Classifier, the Random Forest, and the Logistic Regression, as well as the development of the Ensemble Method (Bagging & Boosting), which uses classifiers like the XGBClassifier and the Bagging Classifier to combine various amounts of classification models for identification.
Keywords:- Detection Fake News, Scraping, Social Media, Text classification, Comparison of Algorithms, Machine learning, Natural language process.
