We manually annotated various eye-tracking corpora with named entity labels (PERSON, ORGANIZATION, LOCATION). We do not re-distribute the original eye-tracking data. Hence, these files only contain text and named entity annotations.
This repository contains the data used for the following papers:
"Entity Recognition at First Sight: Improving NER with Eye Movement Information" by Nora Hollenstein & Ce Zhang (NAACL, 2019).
Data used: Dundee, GECO, ZuCo 1.0 (sentiment and normal reading)
"Advancing NLP with Cognitive Language Processing Signals" by Nora Hollenstein et al. (ArXiV preprint, 2019).
Data used: ZuCo 1.0 (sentiment and normal reading)
The datasets were annotated by two NLP experts. The IOB tagging scheme was used for the labeling. We followed the ACE Annotation Guidelines (Linguistic Data Consortium, 2005).