| Abstract : | People that work with data, store it by means of databases, xml files, worksheets or even flat files. A challenge arising here is how people can easily exchange data, since there is high heterogeneity and eventually how their data relate to each other. The field of data integration deals with the above mentioned problems with two separate approaches. Warehousing is the first approach, where data is gathered from the sources and translated to a target warehouse database with a process called ETL (Extract, Transform, Load) or by utilizing declarative formulae called TGDs (Tuple Generating Dependencies).The second one is Virtual Data Integration where data are accessed through a virtual database called mediated schema. Recent research focus on how to use ontologies as the mediated schema for VDI purposes, to ultimately gain better interoperability. Ontologies are a way to conceptualize a certain domain with better expressivity than other means of data modeling, and potentially act as a unifying framework for heterogeneous data sources, if clients map their local data to a widely accepted ontology. A question that occurs here is how clients can efficiently, accomplish this task of mapping. Our aim is to offer an easy ontology based, data integration system by extending the MIPMap tool with ontology related features.
|
|---|
Copyright © 2013 Library AUEB. All rights reserved.