[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: VMs: Strange pattern (key?)
At 07:55 18/11/2003 -0800, Rene Zandbergen wrote:
Several people have experimented with hidden Markov
models and from talking to one of them (Jim Reeds)
I remember that, when applied to language
(I guess it may have been Latin or English), it
will highlight the alternation between vowels
and consonants. This did not happen with the VMs.
Instead, there it highlighted the well-known word
patterns, e.g. (q)ok(e)(e)(d)y.
Given that I suspect VMs words to be bimodal (ie, with a <qokeedy> mode and
a non-<qokeedy> mode), my current effort is going into consiously
constructing a strongly predictive Markov model for the <qokeedy> mode, and
only then perhaps trying to see if there's another HMM hiding in the
residue. Most people focus on <qokeedy> words, but the remainder may
actually prove far more useful to us, you never know... :-)
Cheers, .....Nick Pelling.....
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying: