RE: RE: Markov algorithms |
||
 Mar Xenographic Sounds cool, though I can't say that I've really delt with Modula-2 much... :] What I really need are descriptions of the algorithms. A basic Markov chain itself isn't that hard--the Perl implementation I have is *DEAD* easy... It's just a hash with elements of $keylen chars and the data for that key is a list of possible next chars (which I later sort). Sure, that's pretty efficient, but I can always gzip the chains if they get too big, or replace the data with an object if I get chains *THAT* big... :] I think that part of my problem is that I'm trying to connect many different chains together, rather than generating them off just one source & I just have to find some way to see to it that I tie all the ends together myself (which is a PITA, but not impossible, so long as I guarantee that the chains cannot have "ends" on them, as it were). Of course, it's the tight loop problem that really bugs me & which I can't stop quite as easily. It's just annoying when otherwise wonderful output is too repetitive :] At some level, I can't totally stop it & I just need a lot more input, but the smaller loops are still a pain... :] Pity I can't generate a graph of the keys... I'm sure that they'd look like a terrible squiggle, but... :] Replies:
|
||
| CyberArmy::Forum v0.6 Generated In 0.00557 seconds |
2002-10-29 11:16:29
2002-10-29 08:24:32