Skip to content
Branch: master
Go to file

Latest commit


Failed to load latest commit information.
Latest commit message
Commit time

Entity Linking integrated Retrieval (ELR)

This repository contains resources developed within the following paper:

F. Hasibi, K. Balog, and S.E. Bratsberg. “Exploiting Entity Linking in Queries for Entity Retrieval”,
In proceedings of ACM SIGIR International Conference on the Theory of Information Retrieval (ICTIR ’16), Newark, DE, USA, Sep 2016.

You can check the paper and presentation for detailed information.

The repository is structured as follows:

  • nordlys/: Code required for running entity retrieval methods.
  • data/: Query set and data required for running the code.
  • qrels/: Qrels files for the DBpedia-entity test collection (version 3.9).
  • runs/: Run files reported in the paper.


Use the following command to run the code:

python -m nordlys.elr.retrieval_elr <model_name>

Using this command, the retrieval results are produced using the recommended parameters in the paper. For detailed descriptions and setting different parameters read the help using the command python -m nordlys.elr.retrieval_elr -h.

Python v2.7 is required for running the code.


Check the nordlys/elr/ file for the actual implementation of the ELR framework and the baseline methods.


The indices required for running this code are described in the paper. You can also contact the authors to get the indices. The following files under the data folder are also required for running the code:

  • queries.json: The DBpedia-entity queries, stopped as described in the paper.
  • tagme_annotations.json: Entity annotations of the queries obtained from the TAGME API.


If you use the resources presented in this repository, please cite:

   author =    {Hasibi, Faegheh and Balog, Krisztian and Bratsberg, Svein Erik},
   title =     {Exploiting Entity Linking in Queries for Entity Retrieval},
   booktitle = {Proceedings of ACM SIGIR International Conference on the Theory of Information Retrieval},
   series =    {ICTIR '16},
   year =      {2016},
   pages=      {209-218},
   publisher = {ACM},
   DOI =       {ttp://}


Should you have any questions, please contact Faegheh Hasibi at


Exploiting entity linking in queries for entity retrieval




No releases published


You can’t perform that action at this time.