Deriving and Validating Emotional Dimensions from Textual Data

This paper analyses information extraction methodology of underlying emotional dimensions connected to different textual data, including social-media posts and online reviews. By combining previously academic work on emotional dimensions, basic emotions and emotional sentiment we analyse the intersection of all three areas by using principal component analysis for information retrieval of orthogonal dimensions with highest variance in emotional variability. We identify that our results follow a coherent conclusion across all 16 studied datasets. In particular, the found orthogonal emotional dimensions are a combination of valence (positive-negative sentiment), activation arousal (arousal-dominance), and expectancy tension (intensity towards future). We are able to confirm the existence of both valence and arousal as core dimensions, consistent with some newer academic papers dominance does not emerge as an individual component, but it appears as attribute connected to the variability of both valence and activation arousal dimensions. We also find evidence of existence of "unpredictability/novelty” dimension from recent academic research. We propose for key scientific contribution that empirical results show that an additional orthogonal emotional dimension should be defined and named "expectancy tension" dimension in that it captures brain activation and emotional variability linked to the intensity regarding future scenarios. Besides the aforementioned results, our work contributes to the social computing literature by suggesting a novel methodology to derive emotional spaces from multiple textual data through eigenvector analyses.