Cross-lingual ontology alignment (COAL)

This tool is useful for aligning ontologies in two different languages

The structure of the code is as follows

• main: The aligner
• nlp.core: Basic NLP methods (tokenization, etc.)
• nlp.sim: Cross-lingual similarity metrics
• test: Integration tests
• translation: Interface to the Monnet Translate system
• web: Web interface for COAL

Installing and Running

COAL can be installed using Maven simply as follows

mvn install

Running the aligner can be done with ./main/align script as follows

./align Ontology1.owl Ontology2.owl Alignment.rdf

Example ontologies can be found at main/load/SourceOntology.owl and at main/load/TargetOntology.owl.

In addition a you can require COAL to provide more matches as follows:

./align Ontology1.owl Ontology2.owl Alignment.rdf 5

COAL can be trained as follows

./train Ontology1.owl Ontology2.owl Alignments.rdf modelOutFile

COAL is built on SVM-Rank and so will only work on Linux and Windows

Web server

The web server can be installed by compiling and uploading the WAR file found under web/target to a Java EE container.