A SEMANTIC FAKE NEWS DETECTION SYSTEM USING MACHINE LEARNING CLASSIFIER

Abstract

The purpose of fake news detection system is to build ontology to find hypothesis involved in misleading social media users through automated reasoning. Ontology for classification of news content has been created after understanding the semantic notations of textual features with in fake news dataset. The dataset we have used in our approach openly available on open-source data repository with the name fake News. The proposed model will provide semantic analysis of news content of the dataset and classification of news content into fake categories. The outcome of our proposed solution can be originating by applying three different classifiers of machine learning that is Random Forest, Logistic regression and LSTM (Long Short-Term Memory) that showed results about fake news and the accuracy of our proposed methodology is almost about 97%, 98% and 99% respectively. Thus the results prove that machine learning models performed better after analyzing the semantic features from news datasets.