helvirago | Project, baby, Project

Hey! Welcome back, fellow masochists! As a special treat, today we're going to talk about

Methods

and more specifically,

1. Cusum

(Hmm. I'm being distracted by my work music. Man, I really love SOL's "Blue Tattoo" on the Hard Core Logo tribute album. Anybody else know what I'm talking about?)

This part I already wrote, for the part of the project that I thought I was going to turn in earlier this semester, but then she didn't look at it -- so that was sort of a waste of my time, but on the other hand, score! it's already written!.

The cusum (or Qsum) technique relies on the theory that the interaction between certain stylistic habits -- for example, sentence length and number of function words per sentence -- is particular (UNIQUE? IDIOSYNCRATIC? WORD?) and consistent for one author. I used sentence length and the combination of 2- and 3-letter words and words beginning with a vowel per sentence, as this is reported to be the pair of attributes most useful for author tracking (Farringdon).

For a section of text to be analyzed, the mean for each statistic is calculated. Then the deviation from the mean is calculated for each sentence, and the cumulative sum is graphed. Because these stylistic habits are unconscious and not directly related to the topic of the writing, the relationship between the two should (according to the method's proponents) remain consistent for the same author. To test the authorship of a piece of text, the "unknown" piece is inserted in the middle of work known to be by a candidate author, and the cumulative sums for the whole are computed. If the relation between the two running sums remains roughly constant throughout the whole piece (known and unknown), the unknown text can be attributed to the author. If the unknown piece is by a different author, the relationship between the two (scaled) will be uneven, and the lines will converge or diverge.

Should anyone here have read this far and actually wonder what the hell I'm talking about, there's an explanation here.

More once I've had more coffee, and possibly some drugs.

By the way, I found my cat (or rather, she finally decided to come out from wherever she was and find me). All is well, other than the fact that she's so cute she saps my will to do almost anything.