January 13, 2014
Researchers create a computer model to determine a book’s “success”
by Claire Kelley
A team of computer scientists at Stony Brook University have attempted to build statistical models to predict whether or not a book will be “successful.” They define success as the number of downloads on Project Gutenberg, awards won, or Amazon sales. However, they do not distinguish between commercial success and critical success.
By analyzing writing style and linguistic patterns such as sentence structure, balance of parts of speech and increased use of verb tenses (more verbs contribute to better readability), they claim to be able to predict a book’s success with an “accuracy of up to 84 percent.”
In terms of the readability measure, the researchers say there’s an inverse relationship:
In sum, our analysis reveals an intriguing and unexpected observation on the connection between readability and the literary success—that they correlate into the opposite directions…We conjecture that the conceptual complexity of highly successful literary work might require syntactic complexity that goes against readability.
More successful books will also use more plural and proper nouns and possessive pronouns, while less successful books will use more past participle verbs (see Table 6).
Claire Kelley is the Director of Library and Academic Marketing at Melville House.