Integration of web sources under uncertainty and dependencies using probabilistic XML

Abstract

We study in this vision paper the problem of integrating several web data sources under uncertainty and dependencies. We present a concrete application with web sources about objects in the maritime domain where uncertainties and dependencies are omnipresent. Uncertainties are mainly caused by imprecise information trackers and imperfect human knowledge. Dependencies come from the recurrent copying relationships occurring among the sources. We answer the issue of data integration in such a setting by reformulating it as the merge of several uncertain versions of the same global XML document. As an initial result, we put forward a probabilistic XML data integration model by getting some intuitions from the versioning model with uncertain data we proposed in [5]. We explain how this model can be used for materializing the integration outcome. © 2014 Springer-Verlag Berlin Heidelberg.