5 Jun
2008
5 Jun
'08
12:24 a.m.
Hello Andrew, Thursday, June 5, 2008, 12:50:28 AM, you wrote:
However, if you can find me a source that explains what a "Markov chain" actually *is*, I'd be quite interested.
*afaik*, the idea is simple: order-1 Markov chain is just text generated with respect of probability of each char, and order-n is a text generated using probability of each char in context of previous n-1 chars. you just need to gather stats using previous papers :) you can use this idea both to generate new naturally-looking words and to gather stats about words itself and generate naturally-looking sentences -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com