âA Context-based Name Resolution Approach for Semantic Schema Integrationâ

AUTOR(ES)
DATA DE PUBLICAÇÃO

2008

RESUMO

The Semantic Web envisages providing a broad range of services from various application domains on the Web. These services often work in a collaborative way and are based on decision-making tasks, which are performed better if they are able to take into account all the relevant information available. This scenario has given rise to technical developments in information integration, specially addressing data source heterogeneity. A well-known architecture for information integration is the mediator-based approach, which isolates the user from the distributed data sources using an intermediate layer. In an information integration system, a mediator is a software module that uses a single global schema to execute user queries, reformulate them into sub-queries according to the local data source schemas and also, to integrate sub-query results. A schema integration process, therefore, generates the global (mediation) schema by merging individual data source schemas. The main issue in schema integration is the heterogeneity of the underlying data sources. Therefore, semantic disambiguation is indispensable, since, without knowing the meaning of the elements, structural and syntactic methods for integrating schemas are useless. In general, a process for schema integration results in a global schema and a set of inter-schema mappings and usually comprises a number of basic steps such as pre- integration, schema matching and mapping, schema merging and the generation of the final integrated schema. In schema integration, name resolution is the process of dealing with a number of semantic issues in order to determine which real- world entity a given schema element refers to. The semantic information necessary to carry out the name resolution task is commonly provided by generic and domain-specific vocabularies, excluding specific context-bound interpretations. Schema element names can have different meanings depending on the semantic context to which they are related. Thus, the use of contextual knowledge may provide a more accurate semantic interpretation, allowing restrictions to or changes in the meaning of an element name according to a specific semantic context. We propose a context-based name resolution approach for use in schema integration processes. A distinguishing feature of this approach is that it explores and models the contextual information required for name resolution in different steps of a schema integration process. The contextual information is modeled using an ontology-based approach enabling reasoning, reusability and sharing of information. In addition, we propose a basic and extensible schema integration process focused on issues which allow the main requirements concerning name resolution of data source schema elements to be established and investigated. This process has been developed for a mediator-based information integration system, which adopts the GAV approach and uses XML as a common model for data exchange and integration of distributed Web data sources

ASSUNTO(S)

integraÃÃo de esquemas ontologia de contexto informaÃÃes sensitivas a contexto context-sensitive information context ontology name resolution resoluÃÃo de nomes banco de dados semantics in schema integration

ACESSO AO ARTIGO

Documentos Relacionados