The EventStatus Corpus has been described in our EMNLP-2016 paper:

@inproceedings{huang2016events,

 author = {Huang, Ruihong  and Cases, Ignacio  and Jurafsky, Dan  and Condoravdi, Cleo  and Riloff, Ellen},

 booktitle = {Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing (EMNLP)},

 title = {Distinguishing Past, On-going, and Future Events: The EventStatus Corpus},

 year = {2016}

}

This corpus contains 2954 English and 1491 Spanish news stories that were randomly selected from 
the English Gigaword 5th Ed. and Spanish Gigaword 3rd Ed. licensed with LDC. Each of the documents 
contains a predefined civil unrest event keyword and the full annotation was collected in a strict 
two-stage process: sentence-level annotations and event mention/phrase level annotations. 
For more details, please refer to the above paper.

Specifically, we have annotated five temporal status, here's the mapping 
between labels as shown in the corpus and labels as described in the paper:

PA : Past

OG : On-going

FP : Future Planned

FT : Future Alert

FM : Future Possible


and labels at the sentence-level: 

CU : describes at least one CU event

NO : no civil event


In addition, we included the file lists we used in our experiments for tuning and test. 
We performed 10-fold cross-validation in our experiments and we included the specific 
10-fold splits of the test set as well.

for more questions, please email Ruihong Huang at

huangrh@cse.tamu.edu.