Search Results for: Deception Detection In Politics

Politicians use deceit as a successful strategy for political gains and positive image creation. Political fact checking organizations assess the political statements for being truthful or deceitful. This is a challenging problem and has enormous social and political impacts. Automatic deception detection has gained interest in applications, such as security purposes, criminal investigations, and social interactions over the past years, however, there has not been significant statistical analysis for deceit detection in political statements. This gap in analysis can be attributed to the lack of any labeled multimodal dataset for political deception detection. In this thesis, we collected a novel dataset for analyzing political deceit from videos which are labeled by fact checking organizations. The dataset consists of 180 videos from 88 politicians from the two main political parties, including 87 Democratic and 93 Republican videos. An empirical analysis is conducted using the new dataset to investigate political deception detection using linguistic, acoustic and visual modalities. The experimental results indicate the feasibility of detecting political deception using multimodal automated systems as well as specify behavioral patterns that are associated with truthful and deceptive statements from different politicians and different parties.