Thank you Emma. I wasn’t at all bitter since your results supported my thesis — notwithstanding my clumsy “over-fitting” of the first sample (i.e. Lover’s Complaint x 2). However, in a later test I introduced a differently named author (who was still really Shakespeare — I called him Will Shakespeare) and used “The Rape of Lucrece” as the learning sample. My hope was that this re-naming of the author would free this later test from any of your learning from the earlier sample. The comparison text was a letter by Sir Henry Neville which produced an 85% match.
I’m delighted you are working on fine-tuning your ability to analyse historical texts and would be most interested to re-test when you believe the results may be more certain. Alternatively you might like to use the letter of Neville’s (1612) against later text from Shakespeare. I chose Shakespeare poetry rather than dialogue because I felt the character names might confuse things. Shakespeare’s Sonnets (1609) might be usable (having removed the sonnet numbers), though this text is very dense and probably more convoluted for purposed of rhyme and meter than the text of the letter. Best wishes — Bruce
