Detecting Fake News using Support Vector Machines
Bradley Altman, Rachel Kim, Rachel Spear, Jonathan Spychalski
Abstract
In recent years, fake news has emerged as a threat to credible, reliable news. Recent improvements in the capabilities of AI to generate fake news have made it easy to generate realistic but fake news articles that can misinform the public. To help differentiate between real and fake news articles, this paper proposes an SVM model that classifies articles with 89 percent accuracy based only off the title, and a 98 percent accuracy with the title and the first 1000 characters. This high level of accuracy could allow for systems that scan social media or other platforms quickly to flag suspicious articles.
Github Link
Our CodeOur Poster
