We present SchemEX, an approach and tool for web-scale, real-time indexing and schema extraction of Linked Open Data (LOD) at linear runtime complexity. As we cannot assume that a complete retrieval of the LOD cloud on a local machine is feasible, we follow a stream-based approach that makes no assumption about how the RDF triples are retrieved from the web by a data crawler. We show the applicability of our approach by appling SchemEX to the Billion Triple Challenge Dataset 2011 and a smaller dataset with 11M triples.
SchemEX won the Billion Triples Challenge 2011.
For the BTC 2011 SchemEX was computed on two data sets: