From Valence to Emotions: How Coarse versus Fine-Grained Online Sentiment can predict Real-World Outcomes

The growing number of user-generated content that can be found online has led to a huge amount of data that can be used for scientific research. This book investigates the prediction of certain human-related events using valences and emotions expressed in user-generated content with regard to past and current research. First, the theoretical framework of user-generated content and sentiment detection- and classification methods is explained, before empirical literature is categorized into three specific prediction subjects. This is followed by a comprehensive analysis including a comparison of prediction methods, consistency, and limitations with respect to each of the three predictive sources.