The Sem-CrisisLexT26 collection is collection of around 26K crisis-related tweets. The dataset was used in the paper "On Semantics and Deep Learning for Event Detection in Crisis Situations". It extends the CrisisLexT26 dataset by providing concept strings that represent the concepts and entities found in the CrisisLexT26 annotated tweets.
Τhe dataset needs to be aligned with the CrisisLexT26 dataset in order to obtain the informativeness, information type and source labels of each tweet.
- Contents: ~26K tweets posted during 26 crisis events in 2012 and 2013.
- Data format: comma-separated values (.csv) file containing tweet-ids and their concept strings that represent the concepts and endities found in a given tweet.
If you use the Sem-CrisisLexT26 collection, please cite the following paper:
- Burel, G., Saif, H., Fernandez, M., Alani, H.: On semantics and deep learning for event detection in crisis situations. In: Proceedings of the workshop on Semantic Deep Learning (SemDeep) at ESWC 2017 (2017).