How Does Trustium Work?
Trustium’s engine blends the best of Machine Learning and community input to automatically and objectively analyze articles.
When you visit a story, Trustium analyzes the text of the article and identifies the type and categories.
It presents the results with simple icons to let you know if the article is likely credible news, or if it’s questionable and to be avoided. It also indicates if the article is satire, opinion, or something else. Trustium’s analysis examines emotion, bias, and credibility using a variety of indicators.
This is done in real time, so you can quickly make better decisions whether to read an article or not.
Trustium’s Certified Reveiwers are Journalists, English Language Professionals, and other individuals with deep knowledge of news and rhetoric. These experts possesses a wide range of backgrounds and diverse viewpoints. Collectively, they support the training and verification data for the Machine Learning model. This ML engine analyzes each article based on objective evidence rather than any individual’s subjective opinion.
MACHINE LEARNING MODEL
The Machine Learning model uses the training and verification data from the Experts Panel. The system first creates a mathematical model (based on its rhetorical devices and structure) for each article as it is loaded, and then sends that to the Machine Learning model for evaluation. The model responds back in real time with probabilities of how likely that particular article is to be credible news.
Our aim is not to tell you what is true, or even what news to read. Instead, we want to empower you to understand the news you are reading in the crowded and confusing digital environment. Trustium makes it easier to find and read well-written journalistic articles that are unbiased and credible.