I''ve heard that there are methods/algorithms that can be used to determine if a text was written by a certain author (if there is enough material available). I can''t find any good information on the subject, so I was hoping that you could give me a few links, keywords or source suggestions.

The best is to start with :

An algorithm toolbox for on-line cursive script recognition

Edit : hmm..are you looking for this kind of recognition or the "author style" in a non-handwritten text (special sentences, complex words, grammar,..) ?

[edited by - bucheron on September 1, 2003 6:00:32 PM]

quote: Original post by qwert
I've heard that there are methods/algorithms that can be used to determine if a text was written by a certain author (if there is enough material available).

I can't find any good information on the subject, so I was hoping that you could give me a few links, keywords or source suggestions.

Search on things like "author identification" or "author classification". A search on Google (www.google.com) for

Shakespeare author classification PDF

...yielded several pertinent hits (the authorship of some of Shakespeare's works has been the subject of some controversy- there are several other such well-known historical controversies), such as:

http://www.cwi.nl/themes/ins1/publications/docs/SiStDDDM:02.pdf

http://www.umiacs.umd.edu/lamp/pubs/TechReports/LAMP_046/LAMP_046.pdf

http://www.acm.org/sigmod/record/issues/0112/SPECIAL/6.pdf

Personally, I've always been suspicious of this sort of analysis because people can deliberately change their style of writing, and I've wondered whether, statistically speaking, any of these studies includes enough "alternative hypothesis" authors.

Best of luck,
Predictor
http://will.dwinnell.com







[edited by - Predictor on September 2, 2003 6:33:03 AM]

[edited by - predictor on September 2, 2003 10:37:14 AM]

No, I didn''t mean handwriting, although that''s also interesting.

It seems like the support vector machine is one of the most popular methods for handwriting as well as author classification. Do you know where I can find a brief introduction to that and maybe some simple code examples?

Look back over threads form the previous week... there was a thread on Support Vector Machines with plenty of information for you to get started.

Cheers,

Timkin