|Institution||Claude Bernard University of Lyon 1|
|Location||Nautibus (Université Lyon1)|
|remy.delanaux at liris.cnrs.fr|
|Subject||Linked Data integration respectful of confidentiality|
|Abstract||The recent development of Linked Data on the Web enabled the fast publication of huge
amounts of open and linked data, structured as graphs using the RDF format structure;
Integration of data from the Web has to make possible the unified interrogation of various sources, but also the enrichment of an organizational database, the augmentation of applications using content, or even a way to provide a global reference format to publish proprietary data. The absence of a common pattern makes the integration of
Web data rely on instances themselves, matching entities referenced in the nodes of RDF graphs. Yet there may be many references, and they can be erroneous, which prevents a reliable use for integrating the matching data.
Entity resolution is the automated process linking the two identifiers in order to access every information related to this point of interest.
The entity resolution process can reveal confidentiality problems if unexpected linkes between open data and proprietary data are created. The goal of this PhD thesis is to provide a formal and algorithmical solution to define rules for entity resolution and check the compatibility of these rules with security and confidentiality
Scientific publications are considered, along with collaborations for future calls for projects regarding new challenges in Europe and in France on Big Data. The city of Grenoble already committed to provide the data from its Open Data portal for this project.
Last update : 2017-03-01 11:23:22