A similarity measurement tool for OWL 2 ontologies
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
.settings
src
.classpath
.project
LICENSE
README.md
pom.xml

README.md

GoodOD Ontology Evaluator

Copyright (C) 2011--2012 The University of Rostock.

Installation and Usage Instructions

  • Get a working maven installation to build the evaluator.
  • Install the OntoSim package into your maven repository using mvn install:install-file -DgroupId=org.inrialpes.exmo -DartifactId=OntoSim -Dversion=2.3 -Dpackaging=jar -Dfile=ontosim.jar
  • Install the OntoWrap package from the AlignAPI into your maven repository using mvn install:install-file -DgroupId=org.inrialpes.exmo -DartifactId=OntoWrap -Dversion=4.5 -Dpackaging=jar -Dfile=ontowrap.jar
  • Install the secondstring package into your maven repositry using mvn install:install-file -DgroupId=com.wcohen -DartifactId=secondstring -Dversion=SNAPSHOT -Dpackaging=jar -Dfile=secondstring.jar
  • Run mvn package to build and package the evaluator. The corresponding .jar files are placed in ./target/. The one with the -eval-assembly suffix also contains all dependencies.
  • [Only for GoodOD internal use:] Set the GOODOD_REPO_ROOT environment variable (with setenv/export on *nix or set on Win32, for permanently setting it on Windows, please refer to to this document) to the directory containing the GoodOD svn repository.
  • Run the jar file with java -jar and the test description plist or two owl files as arguments. The -h switch will display further usage instructions.

Scope of Classes from the GoodOD Ontology Evaluator

Some of the classes from the packages de.uni_rostock.goodod.checker and de.uni_rostock.goodod.evaluator are specific to evaluation tasks in the GoodOD project, but can be used for generic uses. The classes from de.uni_rostock.goodod.owl are, however, completely independent of GoodOD specifics and can be reused for other similarity measurement tasks.

Configuration

The file src/main/resources/config.plist contains plain text configuration for the package. Configuration files can also be supplies using the -c command line switch. The relevant configuration values are the following:

  • threadCount: specifies the number of threads to use for normalization and measurement.
  • similarity: The (local) similarity measurement algorithm to use. Available values are:
  • SC: Semantic Cotopy (structural, Dellschaft/Staab 2006)
  • CSC: Common Semantic Cotopy (structural, Dellschaft/Staab 2006)
  • CosineVM: Cosine Vector Measure (lexical, Euzenat et al. 2009)
  • TripleBasedEntitySim: Triple Based Entity similarity (structural, Euzenat et al. 2009)
  • The qualified name of any class implementing the de.uni_rostock.goodod.owl.comparison.Comparator interface.
  • ignoredImports: The import IRIs to ignore for ontologies being compared.
  • normalizers: This subtree contains the configurations for the normalizers. In the 'Basic' normalizer, you can reroute imports of ontologies by specifying the key value pairs. The 'chain' key for the NormalizerChain specifies the names of normalizers to run in turn.
  • measures: Configuration for similarity measurement classes. The only available option here is for triple based entity similarity, where you can choose between three aggregation schemes (MaxCoupling, AverageLinkage, Hausdorff).

License

The software is made available under the GNU General Public License (GPL), version 3 or later (see LICENSE).