My PhD thesis looks at how to provide linked data integration for raw data coming from heterogeneous data sources on the Web.
The work addresses issues raised by heterogeneous data mapping, such as difference in data representation, semantic alignment difficulties and
noise cleaning problems raised by automatic data processing.
The overall objective is to provide a generic and scalable service oriented architecture capable of semantically combining multi-origin heterogeneous data, which are unorganized or not semantically described.
The approach relies on a company scenario to implement this architecture to real-world scale.