Using n-grams to identify time periods of cultural influence

ECU Author/Contributor (non-ECU co-authors, if there are any, appear on document)

Gregory P. Knight (Creator)

Institution

East Carolina University (ECU )

Web Site: http://www.ecu.edu/lib/

Advisor

M. H. N. Tabrizi

Abstract: An author's literary style is influenced by the cultural time period in which the author lives. The author's ideas and the words chosen to express them can help identify the cultural time period that most influenced the author. Ideas are expressed in language through sequences of words called n-grams. Over the past several years Google has been engaged in digitizing millions of books. As part of this endeavor Google has created a database of n-grams extracted from these digitized books and has made the database available to researchers online. This is the first time ever that such an extensive repository of cultural data has been made available. This study develops and tests an original method for utilizing Google's database to identify the cultural time period that most influenced the author of a published work. Several undisputed literary works are examined from which sets of n-grams are extracted and compared against the Google database. The frequency and distribution of n-gram matches allow us to determine the cultural time period that most influenced the author. The method is also tested against several literary works having uncertain or disputed authorship and period of composition. The results suggest that the method developed provides a reasonable approximation of the time period of greatest cultural influence for each book. Unexpectedly the results tend to support conclusions reached by another researcher with regard to prior literary influences on the Ern Malley Poems. In addition they lend support to a well-known alternate theory on the authorship of the Book of Mormon.

Additional Information

Publication

Thesis

Date: 2012

Keywords

Computer science, Linguistics, Sociology, Authorship, Culture, documents, forgery, google, n-gram

Subjects

Computational linguistics

Authorship, Disputed

Style, Literary

Influence (Literary, artistic, etc.)

Google Library Project

Email this document to