Sentimental Matters Predicting Literary Quality with Sentiment Analysis and Stylistic Features

Abstract

The task of predicting reader appreciation or literary quality has been the object of several studies. It remains, however, a challenging problem in quantitative literary analyses and computational linguistics alike, as its definition can vary a lot depending on the genre of literary texts considered, the features adopted, and the annotation system employed. This paper attempts to evaluate the impact on reader appreciation, defined as online users’ ratings, of sentiment range and sentiment arc patterns versus traditional stylometric features. We run our experiments on a corpus of English-language literary fiction, showing that stylometric features alone are helpful in modelling literary quality, but can be outperformed by analysing the novels’ sentimental profile.