• Corpus ID: 211167666

How to quickly detect the overlap and the consistency between LADM with LandInfra and LandXML

@inproceedings{Pouliot2019HowTQ,
  title={How to quickly detect the overlap and the consistency between LADM with LandInfra and LandXML},
  author={Jacynthe Pouliot and C. Monney and Jens Ingensand and Suzie Larriv{\'e}e},
  year={2019}
}

• J. Pouliot, C. Monney, +1 author S. Larrivée
• Published 2019
• Computer Science

In this paper, we explore the schema matching techniques to compare the content of three geospatial standards which are LADM, LandInfra (InfraGML) and LandXML. Those standards all refer to the concept of “land” and we will try to quantify the similarity of them based on syntax and semantic comparison of the class names exposed in their respective schema. Consequently, we will demonstrate the applicability, the accuracy and the usefulness (rapidity and automation) of schema matching techniques for comparing the content of standards. The comparison is performed with XSD (XML Schema Definition) files that describe the schema in English. The results show that syntactic match rate between LADM-LandInfra (54%) is higher than LADM-LandXML (10%). In adding the semantic information extracted from Wordnet, the match rate between LADM-LandInfra goes to 84% and 59% for LADM-LandXML. In comparing our matching results with two independent sources of information that already and manually compared these three standards, we obtained distinctive results. The correctness of LADM-LandInfra is 60%, while the correctness of LADM-LandXML is only 20%. The applicability of schema matching is positively demonstrated while the usefulness and the accuracy still need further improvements in order to make any statement.