Multilingual and cross-domain temporal tagging
Citations
212 citations
188 citations
183 citations
174 citations
Cites methods from "Multilingual and cross-domain tempo..."
...LIMSI (Grouin and Moriceau, 2016) submitted 2 runs for each phase, based on conditional random fields with lexical, morphological, and word cluster features, and the rule-based HeidelTime (Strötgen and Gertz, 2013)....
[...]
148 citations
Cites methods from "Multilingual and cross-domain tempo..."
...LIMSI (Grouin and Moriceau, 2016) submitted 2 runs for each phase, based on conditional random fields with lexical, morphological, and word cluster features, and the rule-based HeidelTime (Strötgen and Gertz, 2013)....
[...]
...LIMSI (Grouin and Moriceau, 2016) submitted 2 runs for each phase, based on conditional random fields with lexical, morphological, and word cluster features, and the rule-based HeidelTime (Strötgen and Gertz, 2013)....
[...]
References
872 citations
"Multilingual and cross-domain tempo..." refers background in this paper
...For example, in topic detection and tracking, it helps to identify new unreported events and to assign documents to already detected events (see, e.g., Allan 2002; Makkonen et al. 2003)....
[...]
797 citations
Additional excerpts
...Although there are some promising machine learning approaches for the extraction of temporal expressions, we developed HeidelTime as a rule-based system for the following reasons: (1) the divergence of temporal expressions is very limited compared to other named entity recognition and normalization tasks, e.g., the number of persons and organizations as well as the variety of names referring to these entities are probably infinite, (2) the normalization is hardly solvable without using rules, (3) resources for additional languages can be added without the need of an annotated corpus, and (4) the knowledge base can be extended in a modular way, e.g., for adding events and their temporal information such as ‘‘soccer world cup final 2010’’ that took place on July 11, 2010. Furthermore, for the ability to easily add and modify rules (req. E), we developed a well-defined rule syntax (see Sect. 4.1.2). As annotation format, HeidelTime uses the TimeML annotation standard of TIMEX3 tags for temporal expressions. Nevertheless, due to the similarities between TIMEX3 and TIMEX2, the tags can be converted into TIMEX2 as well—although not all attributes are supported. Similar to the transformation from TIMEX2 to TIMEX3 described by Saquete Boro (2010), though the other way around, we used this property to be able to evaluate HeidelTime on corpora annotated with TIMEX2....
[...]
392 citations
389 citations
"Multilingual and cross-domain tempo..." refers background or methods in this paper
...On both corpora, HeidelTime significantly Table 3 Results of TempEval2 (Verhagen et al. 2010) and HeidelTime’s publicly available version P R F Value Type...
[...]
...In the context of TempEval-2, we developed HeidelTime’s first version of English resources using the TempEval-2 training data, which corresponds to the TimeBank corpus (Verhagen et al. 2010)....
[...]
332 citations
"Multilingual and cross-domain tempo..." refers methods in this paper
...However, these modifications were not performed using an annotated corpus but in the context of our work on spatio-temporal document exploration (Strötgen and Gertz 2010b)....
[...]
...HeidelTime achieved the best results for both the extraction and the normalization task (English) (Strötgen and Gertz 2010a)....
[...]
...For example, we built a system called TimeTrails for the exploration of events in documents based on the spatial and temporal information occurring together in the sentences of documents....
[...]
...Thus, for our research on multilingual temporal information extraction and exploration (Strötgen et al. 2010; Strötgen and Gertz 2010b), we developed HeidelTime, a temporal tagger satisfying the following requirements: A. Extraction and normalization should be of high quality....
[...]
...Finally, a CAS Consumer writes all extracted pairs of spatial and temporal expressions and thus all events into a database, which is used as knowledge base for the visualization and exploration components of TimeTrails (Strötgen and Gertz 2010b)....
[...]