Authorship Attribution Using Diversity Profiles

AbstractWe propose a new methodology for testing whether two writing samples were written by the same author. While many such tests are based on an index of lexical richness, we propose to use an entire profile of such indices. Specifically, we evaluate a profile of generalized Simpson’s indices for two writing samples and see if the profiles are significantly different or not. We validate our methodology on several poems whose authorship is known. We then apply it to test whether the poem ‘Shall I Die?’ which is sometimes attributed to William Shakespeare was, in fact, written by him. Further, we provide R code and a package for R that easily implements this methodology.