Integration of legal datasets: from meta-model to implementation
Angelo Di Iorio, Gioele Barabucci, Francesco Poggi, Fabio Vitali – {diiorio,barabucc,fpoggi,fabio}

This paper presents ALLOT, an ontology based on Akoma Ntoso to describe entitiesreferenced in legal documents. The focus is on the use of ALLOT to bridge heterogeneous knowledge bases. We present some experiments in integrating heterogeneous legal datasets through ALLOT. We also discuss the lessons learned in the development of this ontology, lessons that can highlight some issues with the current landscape of legal formats and data.

ALLOT is a proof-of-concept ontology we developed based on the Akoma Ntoso non-ontology following the associated guidelines and best practices. The ALLOT ontology is meant to be used to describe in detail the references present in Akoma Ntoso documents, both documental and non-documental references. ALLOT can also be used to bridge KBs extracted from Akoma Ntoso documents to KBs that use other ontologies such as Metalex, PARL (, OCD (, etc.: one of the objectives of this work is indeed to make a first step towards the construction of a Web of Data in the legal field. By integrating information from different KBs it is possible to retrieve complex information on large amounts of data in a single way using Allot.

ALLOT is composed of three layers:

  1. core, where the TLC are declared and documented;
  2. implementation, where the TLC are implemented in terms of well known ontologies such as FOAF or SKOS;
  3. external integration, small ontologies used to align the ALLOT implementation to existing datasets based on other ontologies. Three modules have been alread implemented: a SKOS alignment to Metalex, a SPARQL alignment to OCD and a SPARQL alignment to PARL.

As expected, ALLOT uses the naming conventions discussed in the previous section and it is strongly tied up with the TLC.