[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Fwd: Re: [Gnumed-devel] phrase wheel

From: Hilmar Berger
Subject: Fwd: Re: [Gnumed-devel] phrase wheel
Date: Sun, 14 Sep 2003 15:13:32 +0200 (MEST)

> default_word_separators = re.compile('[- \t=+&:address@hidden')
>  - words inside a phrase must be separated by any combination
>    of at least one of the above
>  - "address@hidden", "left-clavicular", "pain=severe", ...

> default_phrase_separators = re.compile('[;/|]+')
> default_ignored_chars = re.compile("""[?!."'\\(){}\[\]<>~#*$%^]+""")
>  - any that I missed or that should be left out ?
> The parsing then goeth thus:
> - remove ignored_chars
> - split at word_separator boundaries

I'm sorry but I don't understand the concept of these changes. Do you want
to match groups of words / whole phrases ? 
I'm not quite sure if this will work the same way as finding a match for a
single word. In groups / phrases the probability of the match of the whole
term depends on the linked probabilities of each of the parts (like in a
tree model). You will need something like the T9 algorithm used in mobiles
enter words, just that the most probable match changes after each word
typed/matched before.


COMPUTERBILD 15/03: Premium-e-mail-Dienste im Test
1. GMX TopMail - Platz 1 und Testsieger!
2. GMX ProMail - Platz 2 und Preis-Qualit├Ątssieger!
3. Arcor - 4. - 5. T-Online - 6. - 7. daybyday - 8. e-Post

reply via email to

[Prev in Thread] Current Thread [Next in Thread]