Research On Organization Ontology Construction Based On Wikidata

Organization means social entity that has a common goal.As the publisher of information resources,statistical unit for scientific evaluation,important field for information retrieval,representative element of knowledge navigation,organization entity plays an important role in information organization.Organization entitiy is complexity.First of all,there are numerous organization entities,their hierarchical relationship are complex,the forms of entity name are various,and the derivations are diverse.In the context of the booming of LOD,organization entities of different sources are organized in different ways,revealed in different granularities and orientations,presented in different forms,overlapping in coverage.It aggravates the heterogeneity and dispersion of data,makes the uniqueness of the discriminating organization a difficult problem,brings difficulties to the organization-centered information retrieval,bibliometrics,and knowledge navigation,greatly increasing the cost of data mining driven by organization.Ontology is an important tool for maintaining semantics.It can use a defined vocabulary to express links between resources in a standardized and meaningful way and reveal existing and implicit association networks between organization entities.Constructing organization ontology has important theoretical and practical significance for solving data heterogeneity problems,revealing relational networks,discovering potential knowledge,and correctly locating scientific research achievements.Based on Wikidata and DBpedia knowledge base,this paper conducts theoretical and empirical research on the construction of organization ontology.The organization ontology construction process is carried out from the following aspects:(1)The definition of organization ontology attributes and classes.As the basis of constructing the ontology of the organization,the attributes of the Wikidata and DBpedia knowledge bases with a large number of organizational entities are extracted for Attribute Alignment.The Word Netbased attribute alignment method is proposed and implemented.Attributes of the organization ontology is summarized according to the attribute fusion results of the two knowledge bases.Required classes in the ontology are listed by domain and range of the object attributes.In the end,following the principle of vocabulary reuse,the data dictionary of organization ontology is established.(2)Propose the construction method of the taxnomy of the organizational class.The classes associated with the subclass-of attribute and the instance-of attribute in the Wikidata knowledge base is extracted.According to the transitivity of the upper and lower classes,the organizational category tree is constructed.The statistical analysis method is used to analyze the extracted organization category tree,and the optimization framework is designed for correcting existing problems.The optimized organization category tree is used as classification for organization in the ontology.(3)Using the Protégé to formalize the ontology model,adding instances based on the ontology model.Combining with the backward compatibility of OWL,the data in form of tables are transformed into the data in semantic RDF format to realize the batch import of organization instances and the improvement of the knowledge base of organization ontology,so as to verify the validity of institutional ontology.