Pairwise link discovery approaches for the Web of Data do not scale to many sources thereby limiting the potential for data integration. We thus propose a holistic approach for linking many data sources based on a clustering of entities representing the same real-world object. Our clustering approach utilizes existing links and can deal with entities of different semantic types. The approach is able to identify errors in existing links and can find numerous additional links. An initial evaluation on real-world linked data shows the effectiveness of the proposed holistic entity matching.

<h2 id="bibtex_heading">BibTex</h2>
<pre id="bibtex_listing">

@inproceedings{Nentwig:ICDM2016,
title = {Holistic Entity Clustering for Linked Data},
author = {Markus Nentwig and Anika Groß and Erhard Rahm},
booktitle = {{IEEE} International Conference on Data Mining Workshop, {ICDMW} 2016,
Barcelona, Catalonia, Spain, December 12-15, 2016},
year = {2016},
crossref = {DBLP:conf/icdm/2016w}
}

@proceedings{DBLP:conf/icdm/2016w,
title = {{IEEE} International Conference on Data Mining Workshop, {ICDMW} 2016,
Barcelona, Catalonia, Spain, December 12-15, 2016},
publisher = {{IEEE} Computer Society},
year = {2016}
}

</pre>