Hi Helen,

UMLS::Similarity defaults to using MeSH (MSH). I don't think both of those
concepts are in Mesh, so that's why you get -1 (as do I when I test it
here). However, if you use SNOMEDCT as your source then you get the result.
Most of our experimental results (that you see in papers) have been based
on using SNOMEDCT.

Below see my command to use SNOMEDCT...

ted@atlas:~$ umls-similarity.pl --measure wup abortion miscarriage --config
config/snomedct.config
Default Settings:
  --precision 4

User Settings:
  --config config/snomedct.config
  --measure wup


CuiFinder User Options:
   --config option set


UMLS-Interface Configuration Information
  Sources (SAB):
    SNOMEDCT
  Relations (REL):
    CHD
    PAR
  Database:
    umls (MMSYS-2011AA-20110426)




0.8421<>abortion(C0156543)<>miscarriage(C0000786)

Here's what the configuration file looks like....

ted@atlas:~$ cat config/snomedct.config
SAB :: include SNOMEDCT
REL :: include PAR, CHD

And then here is the result with MSH...

ted@atlas:~$ umls-similarity.pl --measure wup abortion miscarriage
Default Settings:
  --precision 4

User Settings:
  --measure wup



UMLS-Interface Configuration Information:
(Default Information - no config file)

  Sources (SAB):
     MSH
  Relations (REL):
     PAR
     CHD

  Sources (SABDEF):
     UMLS_ALL
  Relations (RELDEF):
     UMLS_ALL


-1.0000<>abortion<>miscarriage

I hope this helps!

Good luck,
Ted

On Tue, Apr 24, 2012 at 4:36 PM, helen_geb <[email protected]> wrote:

> **
>
>
> Hi Ted,
> I was trying to find the similarity between words using umls-similarity.pl.
> For most it works fine.
>
> But for some or example abortion<>miscarriage it returns -1. I got those
> words from umls similarity version 0.13 table 1("umls similarity results")
> the output in that version was 0.47.
>
> I don't know how the current version is returning -1? Could it be the
> problem with my data? Or it is returing -1 for every one?
>
> Here is what I am running
>
> >perl umls-similarity.pl --username=root --password='xx" --database=
> umls --measure=wup abortion miscarriage
>
> Thanks,
> Helen
>
>  
>



-- 
Ted Pedersen
http://www.d.umn.edu/~tpederse

Reply via email to