Analytics over Probabilistic Unmerged Duplicates
Ekaterini Ioannou, and Minos Garofalakis
In Proceedings of the 8th Conference on Scalable Uncertainty Management (SUM), September 2014, London.
pdf

Abstract

This paper introduces probabilistic databases with unmerged duplicates (DB^{ud}), i.e., databases containing probabilistic information about instances found to describe the same real-world objects. We discuss the need for efficiently querying such databases and for supporting practical query scenarios that require analytical or summarized information. We also sketch possible methodologies and techniques that would allow performing efficient processing of queries over such probabilistic databases, and especially without the need to materialize the (potentially, huge) collection of all possible deduplication worlds.

Bibtex

@inproceedings{conf/sum/IoannouG14,
     author = {Ekaterini Ioannou and Minos Garofalakis},
     title ={Analytics over Probabilistic Unmerged Duplicates},
     booktitle = {SUM},
     pages = {203-208},
     year = {2014}
}