Citation Reason Dataset

dataset

A dataset of ~4K statements from English Wikipedia annotated with the reason why they need a citation.

Each line of this tab-separated file contains:

* entity_id: the Wikidata ID corresponding to the page

* revision_id: the revision of the corresponding Wikipedia article

* timestamp: the timestamp of the revision

* entity_title: the page/Wikidata ID title

* section_id: the section ID where the statement is

* section: the section title

* prg_idx: the index of the paragraph in the page

* sentence_idx: the index of the statement in the paragraph

* statement: the statement text

* citations: the source cited in the statement:

* vote1: first Mechanical Turk judgment

* vote2: second Mechanical Turk judgment

* vote3: third Mechanical Turk judgment

The numbers in the last 3 fields correspond to the following citation reasons:

1='direct quotation'

2='statistics'

3='controversial'

4='opinion'

5='life'

6='scientific'

7='historical'

8='other'

History