UMLS-Similarity
This package consists of Perl modules along with supporting Perl
programs that implement the semantic similarity and relatedness measures
described by Leacock & Chodorow (1998), Wu & Palmer (1994), Nguyen and
Al-Mubaid (2006), Rada, et. al. 1989, Patwardhan (2003), Jiang & Conrath
(1997), Resnik (1995), Lin (1998), Patwardhan and Pedersen (2006) and a
simple path based measure.
TESTING PLATFORMS
UMLS-Similarity has been developed and tested on Linux and Solaris,
primarily using Perl and the C shell (csh).
REQUIREMENTS
UMLS-Similarity REQUIRES that the following software be installed. More
details on how to obtain and install appear below.
--Programming Languages Perl (version 5.8.5 or better)
--CPAN modules UMLS::Interface
PROGGRAMMING LANGUAGES
Perl (version 5.8.5 or better)
Perl is freely available at:
<http://www.perl.org>
It is very likely that you will already have Perl installed if you are
using a Unix/Linux based system.
CPAN MODULES
UMLS-Interface
Please see the UMLS-Interface documentation.
INSTALLATION
If you have supervisor access, or have configured MCPAN for local
install, you can install via:
perl -MCPAN -e shell
> install UMLS-Similarity
If not, you can, "manually" install by downloading the *.tar.gz file,
unpacking, and executing the following commands.
perl Makefile.PL PREFIX=/home/programs LIB=/home/lib
make
make test
make install
Note that the PREFIX and LIB settings are just examples to help you
create a local install, if you do not have supervisor (su) access.
You must include /home/lib in your PERL5LIB environment variable to
access this module when running.
If you have the super-user access, then you can install UMLS-Similarity
into system directories via : perl Makefile.PL make make install make
clean
The exact location of where UMLS-Similarity will be installed depends on
your system configuration. A message will be printed out after 'make
install' telling your exactly where it was installed.
USING vector.pm
The vector measure is in a pre-beta stage at this point. It requires the
BerkelyDB which I apologize because after installing the UMLS into MySQL
the last thing you really want to do is install another database. This
shouldn't be as painful and I promise to work on getting this updated so
you will no longer be required to do this.
You can download BerkleyDB from here:
<http://www.oracle.com/technology/software/products/berkeley-db/index.ht
ml>
You can either chose the encripted or non-encripted version. I chose the
non-ensripted because it was smaller. I am using linux (ubuntu to be
specific) so I downloaded the following file:
Berkeley DB 4.7.25NC.tar.gz
Extract the package using gunzip and tar
% gunzip db-4.7.25.NC.tar.gz
% tar -xf db-4.7.25.NC.tar
In the directory, there is a file called doc/index.html which has the
installation instructions. I installed it using the following steps and
series of commands:
% cd build_unix
% ../dist/configure
% make
% sudo make install
Next, download the BerkleyDB.pm module from CPAN:
<http://search.cpan.org/dist/BerkeleyDB/>
The installation directions are in the README. Below are the steps and
commands that I used to install the package.
Extract the package using gunzip and tar % gunzip BerkeleyDB-0.38.tar.gz
% tar -xf BerkeleyDB-0.38.tar
In the BerkeleyDB-0.38 directory, you are going to have to modify the
config.in file.
First, change the path of the INCLUDE variable to point to the directory
where db.h is installed. For example:
INCLUDE = /home/bridget/programs/db-4.7.25.NC/build_unix
Second, change the path of the LIB veriable to point to the directory
where the libdb is installed. For example:
LIB = /home/bridget/programs/db-4.7.25.NC/build_unix
Then continue on with the typical perl module install:
% perl Makefile.PL
% make
% sudo make install
Theoretically, that should do it. As I said before this is the beta
version so in the near future these additional steps will not be
necessary to use the vector measure module.
To use this method you will need to specify the --dbfile option, which
is BerkleyDB file that contains the vector information. If you need a
test file for this please email me at: bthomson@cs.umn.edu
CONTACT US
If you have any trouble installing and using UMLS-Interface, please
contact us via the users mailing list :
umls-similarity@yahoogroups.com
You can join this group by going to:
<http://tech.groups.yahoo.com/group/umls-similarity/>
You may also contact us directly if you prefer :
Bridget T. McInnes: bthomson at cs.umn.edu
Ted Pedersen : tpederse at d.umn.edu