Deceptive Content Analysis using Deep Learning

Abstract

Fake news is the deliberate spread of false or misleading information through traditional and social media for political or financial gain. The impact of fake news can be significant, causing harm to individuals and organizations and undermining trust in legitimate news sources. Detecting fake news is crucial to promote a well-informed society and protect against the harmful effects. Tools such as machine learning and natural language processing are being developed to help identify fake news automatically. Necessity of fake news detection is very important to maintain a trustworthy and responsible media environment. We have used Word2Vec model for word vectorization and represents words in a multi- dimensional space based on their semantic and syntactic relationships. The use of the LSTM with 256 units allows our model to capture the sequential nature of the data and make predictions based on past information. The proposed model uses Word2Vec and LSTM models to provide a powerful approach to fake news detection, combining the ability to capture the complexity of language and the sequential nature of the data. The model has the potential to accurately detect fake news and promote a well-informed society. The accuracy achieved by building the model was 97%.