Opinion mining and lexical affect sensing

The dissertation presents an interdisciplinary study combining linguistic and psychological insights to categorize opinions and emotions in texts through computer-aided methods. It examines various emotional corpora, including movie reviews and weblogs, and outlines different classification approaches: statistical, semantic, and a hybrid of both. Additionally, it explores affect sensing through multimodal fusion of lexical and acoustic data. The work concludes with a discussion of its contributions and potential future research directions.