Authorship of 2 Henry IV

UK-Flagge auf deutsch All data were generated from R Stylo (see: Computational Stylistics Group Homepage).

So far Rolling Delta investigations have checked word frequencies, character bi- and trigrams and looked into
window sizes between 1000 to 5000 words. There were up to 12 reference texts that R Stylo was able to present
graphically. Normally an author was represented by just one or perhaps two reference texts, but a case where a
better suited rarer text produced even lower delta values was not considered.
A new approach makes use of all available reference texts to find the best suited that would have been overlooked.
A normal PC takes about four hours to go through all files with a 5000-word window, a step size of 250 Words and
character trigrams as variables. All window measurements then go into a spreadsheet and the lowest three values of
each measurement are highlighted with conditional formatting. Only the highlighted texts are kept and displayed
in the chart below.
The values of the first measurement of the 5000-word window are shown at 2500 words in line 12. The lowest is de-
picted in green, the next in yellow, and the third-lowest is in red. The sequence of green cells
indicates the most likely author of those windows.



Compare these evaluations with the results of Rolling Classify.