> Drago asked me to respond to this email to the group.
>
> The best advice I can give is to read the README file
>which should have come with the cidr download from
>http://www.summarization.com/mead/addons/index.html The
>basic problem I see from the message above is that you
>can't just execute "perl cidr.pl" At a minimum you need
>to feed it a file which contains pathways to all the
>documents to be clustered. There are also several other
>optional variables which are described in the README
>along with a sample execution.
>
> Thanks.
>
>
> Ali Hakim
yes, i have wrong in execution way, sorry ^^;
now i execute that by "perl cidr.pl all"
file all contains all articles in xml format.
in file readme for using cidr, it say "cidr.pl requires
that the files nidf.dir and nidf.pag be present in the
directory where cidr.pl is executed."
but i don't know how to create nidf.dir and nidf.pag
my linux (redhat9) execute "dbmopen(%nidf, "nidf", 0666);"
just create file "nidf" (just 1 file without *dir or *pag
extention), therefore "$nidf{$key}" always have no value
and similarity become zero and always make a new cluster.
same problem with "dbmopen(%centroid, "centroid", 0666);"
just create 1 file "centroid", so the next error is "mv
centroid.dir centroid.pag T-CLUSTERxxxxx" because file
"centroid.dir" and "centroid.pag" not exist.
what must i do ?
pls reply soon ... thx alot
odil
========================================================================================
Akses Internet TELKOMNet-Instan beri Diskon s.d. 50 % khusus untuk wilayah Jawa Timur.
Informasi selengkapnya di www.telkomnetinstan.com atau hub 0800-1-INSTAN (467826)
========================================================================================
_______________________________________________
Mead mailing list
Mead@lists.si.umich.edu
http://lists.si.umich.edu/mailman/listinfo/mead
This archive was generated by hypermail 2b30 : Tue Jun 09 2009 - 05:00:06 EDT