NLP-Based Model for Identifying Misinformation

Abstract

The massive proliferation of digital media has significantly increased the transmission of fake news and disinformation, which affects the confidence of the population and decision-making. The vast quantity of online content can not be managed through manual methods of fact-checking and requires automated answers. This paper introduces an NLP-based machine learning system that can be used to identify fake information in news stories. The system uses text pre-processing, TF-IDF feature extraction and conventional machine learning classifiers such as Support Vector Machine (SVM) and Logistic Regression. The experimental analysis of the proposed model shows that the model can have an accuracy around 91 per cent, which demonstrates that interpretable and lightweight machine learning methods can effectively detect and identify fake information in academic and real-time prototype settings.