Text as Data

APPENDIX AMood State Identification in Text

ORIGINS OF MOOD STATE IDENTIFICATION

Mood state attribution belongs to a class of text analytics solutions that infer author attributes from text fragments taken from writings of a given user. The use of text analytics by biblical scholars to compare, contrast, and attribute authorship to various biblical passages is a longstanding example of this kind of application. In Chapter 4, Document Content and Characterization, we showed how work by Pennebaker was adapted to inferring the gender of a writer. Pennebaker, among others, also uses text analytics approaches to infer psychological characteristics of text authors.

Assessing user sentiment is an important capability, especially for buyer satisfaction ratings, primarily related to product sales and purchases. In addition to assessing the sentiment of a given document there have also been developments to determine the mood of a given document. Sentiment and mood have various meanings both in general discourse and in psychological assessment. Sentiment is a more specific assessment that tends to be oriented towards a given object and brand and is most usually positive, negative, or neutral. Mood, on the other hand, is a more general characterization of the author's general tone or disposition and the gradation of mood is more general and crosses various dimensions.

Mood state analytics evolved in the field of psychological research and psychometrics and has been adapted to use in the ...