Ontology-based clustering in a Peer Data Management System

AUTOR(ES)

Carlos Eduardo Santos Pires

DATA DE PUBLICAÇÃO

2009

RESUMO

Peer Data Management Systems (PDMS) are advanced P2P applications which enable users to transparently query several distributed, heterogeneous, and autonomous data sources. Each peer represents a data source and exports its entire data schema or only a portion of it. Such schema, named exported schema, represents the data to be shared with the other peers of the system and is commonly described by an ontology. The most studied data management issues in PDMS are related to schema mappings and query processing. These issues can be improved if peers are efficiently disposed in the overlay network according to a semantic-based approach. In this context, the notion of semantic community of peers is of great importance since it aims at logically approximating peers with common interests about a specific topic. However, due to the dynamic behavior of peers, the creation and maintenance of semantic communities is a challenging issue in the current stage of development of PDMS. The main goal of this thesis is to propose an ontology-based process to incrementally cluster semantically similar peers that compose communities of a PDMS. In this process, peers are grouped according to their corresponding exported schema (an ontology) and ontology management processes (e.g. matching and summarization) are used to assist peer connection. A PDMS architecture is proposed to facilitate the semantic organization of peers in the overlay network. In order to obtain the semantic similarity between two peer ontologies we propose a global similarity measure as output of an ontology matching process. To optimize ontology matching an automatic process for summarizing ontologies is also proposed. A simulator has been developed resembling the architecture of the PDMS. The proposed ontology management processes have also been developed and included in the simulator. Experimentations of each application in the context of the PDMS as well as the results obtained from these experiments are presented.

ASSUNTO(S)

peer data management systems ciencia da computacao peer-to-peer cluster analysis similarity measure semantic community ontologia ontology summarization ontology matching sistemas de gerenciamento de banco de dados

Documentos Relacionados

• Ontology-based metadata for e-learning content
• Textpresso: An Ontology-Based Information Retrieval and Extraction System for Biological Literature
• UMA INFRA-ESTRUTURA PARA MATCHMAKING BASEADO EM ONTOLOGIAS
• Uma abordagem baseada em ontologias para a interoperabilidade entre agentes heterogêneos
• Onto-Tools: an ensemble of web-accessible, ontology-based tools for the functional design and interpretation of high-throughput gene expression experiments