Entity Relatedness Project

The Entity Relatedness Project (ERP) is an iniciative to find relations and paths between two entities in Knowlegde Databases.

Dependencies

Lua 5.2
Lua Rocks 2.4.1
lpeg 1.0.0-1
lua-curl 0.3.5-1
luaexpat 1.3.0-1
luafilesystem 1.6.3-2
luasocket 3.0rc1-2
luasql-mysql 2.3.0-1
redis-lua 2.0.4-1
uuid 0.2-1
wsapi 1.6.1-1
GNU parallel 20161122
Redis Server 3.2.6

Architecture Configuration

Ontology Class

At the data pre-processing layer, ERP builds an index (using Redis) over the DBpedia and Wikidata class hierarchy to help identify the immediate classes of an entity. We follow the approach published at [1]. The Onotlogy of DBpedia is available at [2] and the Wikidata Taxanomy is found at the link [3]. In a next version we available this data as a service.

Processing Data

The data extraction is executed throug HTTP requests. We used luasockets and luaexpat for this task. The generated jobs are executed in paralllel with GNU parallel [4]. The URIs indentified in an Entity Document is sored and encoded in a Local Redis Server.

Finding Path

References

[1] Herrera, J., Casanova, M.A., Nunes, B.P., Lopes, G.R., and Leme, L.A. DBpedia Profiler Tool: Profiling the Connectivity of Entity Pairs in DBpedia. Proc. 5th Workshop on Intelligent Exploration of Semantic Data (October 2016).
[2] DBpedia Ontology Ontology.http://downloads.dbpedia.org/2014/dbpedia_2014.owl.bz2
[3] Wikidata Taxonomy: https://tools.wmflabs.org/wikidata-exports/rdf/exports/20160801/wikidata-taxonomy.nt.gz
[4] GNU Parallel: https://www.gnu.org/software/parallel/