PROJECT TITLE :

A Multifaceted Approach to Social Multimedia-Based Prediction of Elections

ABSTRACT:

Compared with real-world polling, election prediction based on social media will be way additional timely and value-effective thanks to the immediate availability of fast evolving Internet contents. However, information from social media may suffer from noise and sampling bias that are caused by varied factors and therefore pose one among biggest challenges in social media-based information analytics. This paper presents a brand new model, named competitive vector auto regression (CVAR), to create a reliable forecasting system for the US presidential elections and US House race. Our CVAR model is meant to investigate the correlation between image-centric social multimedia and real-world phenomena. By introducing the competition mechanism, CVAR compares the recognition among multiple competing candidates. A lot of importantly , CVAR is ready to combine visual info with textual data from rich and multifaceted social multimedia, which helps extract reliable signals and mitigate sampling bias. Therefore, our proposed system can 1) accurately predict the election outcome, 2) infer the sentiment of the candidate photos shared within the social media communities, and 3) account for the sentiment of viewer comments towards the candidates on the related images. The experiments on the 2012 US presidential election at both national and state levels, as well because the 2014 US House race, have demonstrated the power and promise of the proposed approach.