Profs. Chris Alen Sula and Matt Miller (’13) published a new article in Literary & Linguistic Computing 29(3) on “Citations, Contexts, and Humanistic Discourse: Toward Automatic Extraction and Classification.” The abstract of the paper is below, and the paper is free to download through LCC for the next three months: http://llc.oxfordjournals.org/content/29/3/452.
This paper examines prospects and limitations of citation studies in the humanities. We begin by presenting an overview of bibliometric analysis, noting several barriers to applying this method in the humanities. Following that, we present an experimental tool for extracting and classifying citation contexts in humanities journal articles. This tool reports the bibliographic information about each reference, as well as three features about its context(s): frequency, location-in-document, and polarity. We found that extraction was highly successful (above 85%) for three of the four journals, and statistics for the three citation figures were broadly consistent with previous research. We conclude by noting several limitations of the sentiment classifier and suggesting future areas for refinement.
Work on the project began in summer 2012 and the project was presented at the Digital Humanities 2013 Conference in Lincoln, Nebraska.