Who Wrote the 15th Book of Oz? An Application of Multivariate Analysis to Authorship Attribution

Computational methods in authorship attribution

“…Many studies since that of Mosteller and Wallace have shown the efficacy of function words for authorship attribution in different scenarios (Morton 1978;Burrows 1987;Karlgren & Cutting 1994;Merriam & Matthews 1994;Kessler et al 1997;Argamon et al 1998;Holmes 1998;de Vel et al 2001;Holmes et al 2001aHolmes et al , 2001bBaayen et al 2002;Binongo 2003;Juola & Baayen 2003;Zhao & Zobel 2005;Argamon & Levitan 2005;Koppel et al 2005Koppel et al , 2006a, confirming the hypothesis that different authors tend to have different characteristic patterns of function word use.…”

“…The idea is to visualize the differences between texts written by different authors by projecting high-dimensional word-frequency vectors computed for those text onto the 2-dimensional subspace spanned by the two principal components; if good separation is seen between documents known to be written by different authors, then new texts may be attributed by seeing which authors' comparison documents are closest to them in this space. This method was elaborated on by Binongo and Smith (1999), and has been used to resolve several outstanding authorship problems Binongo 2003;Holmes 2003). A related method is ANOVA, as applied, for example, by Holmes and Forsyth (1995) to the Federalist.…”

Literature Fingerprinting: A New Method for Visual Literary Analysis

“…However, some literature specialists think that Ruth Thompson also wrote the 15th book and that the attribution to Baum was only due to commercial motives to ease the transition from one author to the next without losing sales. See [5] for an interesting analysis on the problem. Authorship attribution has also been named stylometry, because the classification is based on the distinct stylistic traits of a document and is independent of its semantic meaning.…”

“…as the set of words. According to [5], function words have the advantage that writers cannot avoid using them, which means that they can be found in every text and almost every sentence. Furthermore, they have little semantic meaning and are therefore among the words that are least dependent on context.…”

“…We took a list of 52 function words that was also used in [5]. For each text block, a feature vector was calculated by counting the frequency of each of the function words, resulting in a 52-dimensional feature vector.…”