FRanCo - A Ground Truth Corpus for Fact Ranking Evaluation

FRanCo is a ground truth corpus for fact ranking acquired using crowdsourcing.
The corpus is built on a sample of 541 concepts drawn from the English version of DBpedia 3.9.

More details about the sample and corpus statistics can be found in the paper:
T. Bobić, J. Waitelonis and H. Sack,
FRanCo - A Ground Truth Corpus for Fact Ranking Evaluation, Proc. the 1st Int. Workshop on Summarizing and Presenting Entities and Ontologies 2015 (SumPre 2015), co-located with ESWC 2015, CEUR Workshop proceedings, Vol. 1556, 2015.

Download a local copy: FRanCo.zip (7,6 MB)

Dataset files:

|--README.txt
|
|--/aggregated
| |--FRanCo.ttl
|
|--/raw
| |--conceptfacts.tsv
| |--concepts.tsv
| |--step1answers.tsv
| |--step1inputsNER.tsv
| |--step2answers.tsv
| |--users.tsv
|
|--/vocabs
| |--front.ttl
| |--vrank.ttl

* /aggregated/FRanCo.ttl
Aggregated corpus in RDF format, encoded in turtle syntax.

* /raw
Database dumps containing raw information in tab-separated format.

* /vocabs
Vocabularies used to represent the data in FRanCo.ttl (vRank[1] and FRont).

[1] A. Roa-Valverde, A. Thalhammer, I. Toma, M.-A. Sicilia, "Towards a formal model for sharing and reusing ranking computations",
Proc. of the 6th Int. Workshop on Ranking in Databases @ VLDB 2012

Creative Commons License
FRanCo is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.