Sentiment analysis on multipolarized social networks

Abstract

We witness polarized debate in a wide range of relevant topics for the society, such as Politics, Sports and Public Issues. In fact, a significant fraction of debate in social media is concentrated on issues that induce polarization in the society. In this dissertation, we develop contributions on three main directions that analyze and make sense of 'online battles'' fought on social media networks over polarizing topics: We start by (1) measuring the strength of polarization on (online) social networks with respect to a given topic discussion, then, in the context of polarized topics, we (2) offer new methods for processing and interpreting opinions expressed on online polarized debate by uncovering from the social psychology literature well-established social theories that describe how people form their opinions on polarized discussions, and finally we (3) study multipolarized social networks, when more than two viewpoints are in conflict with respect to a topic.