Affect as a proxy for literary mood

We propose to use affect as a proxy for mood in literary texts. In thisstudy, we explore the differences in computationally detecting tone versusdetecting mood. Methodologically we utilize affective word embeddings to lookat the affective distribution in different text segments. We also present asimple yet efficient and effective method of enhancing emotion lexicons to takeboth semantic shift and the domain of the text into account producingreal-world congruent results closely matching both contemporary and modernqualitative analyses.Comment: Submitted to Journal of Data Mining and Digital Humanitie