Doctoral candidate Shadi Rezapour successfully defended her dissertation, "From User-generated Text to Insight; Context-aware Measurement of Social Impacts and Interactions Using Natural Language Processing," on July 20.
Her committee included Associate Professor Jana Diesner (chair); Professor Ted Underwood; Roxana Girju, professor of linguistics; and Karrie Karahalios, professor of computer science.
Abstract: Recent improvements in information and communication technologies contributed to an increasingly globalized and connected world. The digital data that is created as the result of people's online interactions consists of different types of information that can be used to analyze people's beliefs, ideas, and biases. This thesis leverages methods, theories, and views from NLP and social sciences to investigate the manifestations of various attributes and signals, namely social impact, personal values, and moral traits in text. Furthermore, we first present a study at the intersection of review mining and impact assessment and provide a comprehensive discussion on different types of impact that information products can have on people. We then investigate the relationship between principles of morality and stance in text and operationalize moral values to enhance the prediction of social effects. Finally, we extend the theory of structural balance to include direction and utilize emotion and morality to study people's interactions in social networks. Overall, this thesis contributes to the emerging field of “social” NLP and broadens the scope of research in this field by utilizing a combination of novel taxonomies, datasets, and tools to examine user-generated texts and providing comprehensive insights about human language, cultures, and experiences.