Entity Linking in Queries: Efficiency vs. Effectiveness

This repository contains resources developed within the following paper:

F. Hasibi, K. Balog, and S.E. Bratsberg. “Entity Linking in Queries: Efficiency vs. Effectiveness”,
In proceedings of 39th European Conference on Information Retrieval (ECIR ’17), April 2017.

You can check the paper and presentation for detailed information.

The repository is structured as follows:

nordlys/: Code required for running all the experiments
data/: Query set and data required for running the code
qrels/: Qrels files for ERD and Y-ERD test collections
runs/: Run files reported in the paper
scripts/: Run and evaluation scripts

Code

Check the nordlys/erd/ for the implementation of ''candidate entity ranking'' and ''disambiguation'' methods. To generate each of the runs, check scripts/run_scripts.sh.

Python v2.7 is required for running the code.

Data

The index and surface form dictionary required for running this code are described in the paper. You can also contact the authors to get them directly. The following files under the data folder are also required for running the code:

Y-ERD.tsv: The Y-ERD test collection
Trec_beta*.txt: The ERD test collection
fb_dbp_snapshot.txt: Snapshot of proper named entities. This file is the intersection of DBpedia 3.9 and Freebase snapshot of proper named entities provided by the ERD challenge 2014 (fb_snapshot_original.tsv.zip is the original file from the ERD challenge).

Qrels

There are two groups of qrels files, used for 1) ''candidate entity ranking'', and 2) ''disambiguation'' steps. The qrels/*_trec.qrels files belong to the first group, and qrels/*_elq.qrels files belong to the second one.

Citation

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

@inproceedings{Hasibi:2017:ELQ, 
   author =    {Hasibi, Faegheh and Balog, Krisztian and Bratsberg, Svein Erik},
   title =     {Entity Linking in Queries: Efficiency vs. Effectiveness},
   booktitle = {Proceedings of 39th European Conference on Information Retrieval},
   series =    {ECIR '17},
   year =      {2017},
   pages=      {40--53},
   publisher = {Springer},
   DOI =       {https://doi.org/10.1007/978-3-319-56608-5_4}
}

Contact

If you have any questions, feel free to contact Faegheh Hasibi at faegheh.hasibi@ntnu.no.

hasibi / EntityLinkingInQueries-Methods