Moods in Book Reviews: Text Mining Approach

Abstract

Spiteri and Pecoskie (2018) proposed a taxonomy of terms to describe emotion and tone in novels. We tested those terms against 5,144 full-text book reviews from the New York Times Book Review to discover whether the proposed terms were used in published reviews to describe books, and of those terms used, which were most used. The objective of this study is to explore whether emotional contents can be identified by a text mining approach without investing too much time and efforts of information professionals. Findings demonstrate that the terms chosen by Spiteri and Pecoskie are used in professional book reviews, though some may be used in multiple ways, rather than only related to emotional content. Results of this work contribute to a larger scale project of testing machine models of identifying emotional content in books and ultimately being able to create automated media recommendation systems that include emotion as an identifier.

User-centered categorization of mood in fiction

“…For example, psychology studies highlight human emotions, affect (Bartsch and Oliver, 2011), and literature studies look at emotions and tones (Hogan, 2011) together from either writers' or characters' perspectives. Information retrieval studies focus on the users of information who consume different types of media materials, such as books, music, and video games; thus, the mood in these studies (Cho et al, 2021a;Hu, 2010) can refer to either users' moods (how information users feel) or the mood of materials that users seek for.…”

“…However, we note that mood is not an independent phenomenon. As an example, contextual situation-wise, readers' current work situations, personal lives, where they are, or even the weather outside can all impact readers' moods, which can influence their reading choice (Cho et al, 2021a). Also, particular scenes or descriptions about places in the fiction book can lead readers to feel certain ways.…”

“…Mood is a deep and complex notion. It refers to various concepts across different domains, such as affect in psychology (Bartsch and Oliver, 2011), emotions and tones in literature (Hogan, 2011), and mood in music information retrieval (MIR) as well as other user studies in information science (Cho et al. , 2021a; Hu, 2010).…”

“…Mood is a deep and complex notion. It refers to various concepts across different domains, such as affect in psychology (Bartsch and Oliver, 2011), emotions and tones in literature (Hogan, 2011), and mood in music information retrieval (MIR) as well as other user studies in information science (Cho et al, 2021a;Hu, 2010). Mood has been understood in various ways, but even with the variances in specific meaning across different domains, existing literature agrees on one thing: Mood is an important element for media information users, especially for the ones who seek leisure materials (Hogan, 2011;Vorderer and Reinecke, 2015).…”