Quote:
Originally Posted by Daithi
I just posted your quote and it came back P. G. Wodehouse for me as well, so it looks like whatever the site is doing is based on some kind of examination of the text.
It would be nice if they described how the algorithm worked. I could see developing something that looked at average sentence length, number of adjectives per sentence, length of words, number of adjectives, adverbs, etc. Actually this would make a pretty good project for a neural network type of program that you train.
I tried entering a short quote from a Proust book and the website guessed Lovecraft. However, I entered something from Frankenstein and it guessed Shelley.
|
Quote:
Originally Posted by Bikeridr
I did the same with beppe's text, and it also came back as Chuck Palahniuk, so it seems that, as you suggest, they are using an algorithm of some sort.
So, lets see how consistent we are. I'll make another text tomorrow (night now in Europe) and see whom I am then.
|
The algorithm is short and sweet. They count words and punctuation marks. They have a very limited base and project the text on it. Pattern recognition.