TimePro is a time expression recognizer that annotates both absolute time expressions (e.g. “January 2017”) and relative time expressions (e.g. “tomorrow”) in a text.
Temporal expressions are first classified (e.g. dates, durations, times) and then normalized according to the ISO-TimeML specifications.
Recognition is based on a statistical classifier, while normalization is rule-based. TimePro is available for Italian and English with pre-trained models in the news domain.
Algorithm: TimePro uses Yamcha for feature extraction and SVM as a classification algorithm.
Resources: EVENTI (Italian) and TempEval 3 (English).
Evaluation benchmark: EVENTI at Evalita 2014 and QA-TempEval 2015 (English).
Paramita Mirza and Anne-Lyse Minard. FBK-HLT-time: A complete Italian Temporal Processing System for EVENTI-Evalita 2014. In Proceedings of the First Italian Conference on Computational Linguistics CLiC-it 2014 & the Fourth International Workshop EVALITA 2014, Pisa (Italy), 11th December 2014.
Paramita Mirza and Anne-Lyse Minard. HLT-FBK: A complete Temporal Processing System for QA TempEval. In Proceedings of the 9th International Workshop on Semantic Evaluation (SemEval 2015), Denver (Colorado, USA), June 4-5, 2015.