ENHANCING VERACITY : LEVERAGING MACHINE LEARNING ENSEMBLE METHODS FOR FAKE NEWS DETECTION

Abstract

The emergence of the World Wide Web and the use of social media platforms such as Facebook, Twitter, and Instagram have led to the development of a technique of disseminating information that was not possible before the digital age. Many of the information available on these social media platforms could be false. As such, it is imperative to keep an eye on this data. It is possible to employ a method known as machine learning-based fake news identification to assess the authenticity of fresh articles or facts by feeding them into the model. Before training the dataset, we will preprocess the data (text in this case). The majority of preprocessing involves removing unnecessary data. After that, the dataset is split into two parts: training and testing. Next, the TF-IDF vectorization approach will be used to vectorize the data. The vectorized data is then used to train the different classifiers (like random forest, svm, xgboost, etc.). These findings are then integrated into ensemble models to improve the precision of state-of-the-art false news detection. The timeliness of a dataset affects the model's accuracy since it prevents the model from accurately predicting the authenticity of more recent information because it excludes information that is too old from previous datasets. The model can be promptly tested using the testing dataset after training, at which point it can be put to use.