Research Of Ontology Mapping Based On Relational Database

The amount of today's Web content outpaces technological progress. Anapproach that can improve the current situation is to represent Web content in aform that is more easily machine-processable, and to use intelligent techniques totake advantage of these representations. We refer to this plan as the SemanticWeb.An ontology is an explicit and formal specification of a conceptualization. Ingeneral, an ontology describes formally a domain of discourse. Typically, anontology consists of a finite list of terms and the relationships between theseterms. The ontology layer represents the meaning of the information and therelations of all kinds of semantic information. The ontology layer is the key of theSemantic Web technology and became a hotspot of the research interest. Thispaper makes a deep investigation about the representation, storage and mappingof ontology. Based on this investigation we propose a way of mapping ontologybased on relational database.There are many different description languages for ontology. At present, themost important ontology languages for the Web is the W3CRecommendation—Web Ontology Language (OWL). It is a richer vocabularydescription language for describing properties and classes, such as relationsbetween classes, cardinality, equality, richer typing of properties, characteristicsof properties, and enumerated classes. Ontology have several storage format:plain text is explicit to represent the semantic of ontologies, and is applied tostore small ontologies;special manage tool is not mature nowadays;rationaldatabase which is more mature suits for large data storing efficiently andconveniently. At present, some popular technique for storing ontologies based onrelation database, for example, horizontal, vertical and decompose pattern are toosimple to preserve the semantic of ontologies sufficiently. They need to connecttables or modify table pattern when it's time to query or evolve ontologies. Thispaper proposes an approach which has some tables to store the resource and theirrelations in owl documents based on relational database to solve these questions.In the context of the Web, ontologies provide a shared understanding of adomain. Such a shared understanding is necessary to overcome differences interminology. Such differences can be overcome by mapping the particularterminology to a shared ontology or by defining direct mappings between theontologies. Mapping is a formal expression that states the semantic relationbetween two entities belonging to different ontologies. Solving the ontologymapping is the first step to semantic interoperability. Previous researches haveproposed many approaches and techniques for specific application domains. Wepresent a taxonomy that covers many of these existing approaches, and describethe schema-level and instance-level, element-level and structure-level, andlanguage-based and constraint-based match approaches.The first step of Ontology Mapping is normalization--unifying all data tothe same representation level, coping with syntactical, structural and languageheterogeneity. The above storage schema normalizes both ontologies to a uniformrepresentation: described by OWL and stored in the relational database. The nextwork is to find mappings. This paper processes schema-level mapping withoutconsidering the instance-level information. We adopt the composite mapping thatcombines the results of several independently executed matchers, such aselement-level and structure-level mapping. First, we propose a linguisticsimilarity algorithm, which compares two concept names using edit distance, andget the primary result. At the structure-level we may compare semantic structures(hierarchical strcture and semantic relation) of two ontologies. In our model theclass context is constituted by the relevant elements (properties and neighbourclass) and their semantic relation. We propose a structure similarity algorithm tomatch two concepts' context. Then, we integrate the results according to thedefinite weight and determine the final mapping relation. The storage method isproved feasible and efficient for retrieval and query. The mapping algorithm canmatch the related concepts automatically and achieve efficient mappings betweentwo ontologies.The method and algorithm for ontology storage and mapping that this paperproposes has an important practical significance. At the end of the article, theauthor looks forward to the applied foreground of ontology, and puts forward tothe points of the future work.