Models of Trust and Disinformation in the Open Press from Model-Driven Linguistic Pattern Analysis

The TRAQ-M (tracking analysis, quantification-mitigation) platform is a computational system that applies Social Science Models and nonparametric statistical methods to understand complex human behavior patterns. The system focuses on understanding patterns in language and is capable of ingesting millions of documents per day and identifying linguistic patterns. In this paper we focus on source modeling, especially determination of which sources pass false information or are otherwise biased We use nonparametric statistical models to compare document content histograms and linguistic pattern analysis to identity disinformation.