Emotional book classification from book blurbs

Knowing and predicting opinions of people is considered a strategic added value, interpreting the qualia i.e., the subjective nature of emotional content. The aim of this work is to study the feasibility of an emotion recognition and automated classification of books according to emotional tags, by means of a lexical and semantic analysis of book blurbs. A supervised learning approach is used to determine if a correlation exists between the characteristics of a book blurb and emotional icons associated to the book by users. In this paper the underlying idea of the system is presented, the preprocessing and features extraction phases are described and experimental results on the social network Zazie and its mood tags are discussed.