--- title: "Temporal Tagging on Different Domains — Strötgen 2012" description: "Most temporal taggers and annotated corpora come from news, and earlier work showed that moving to other text types, such as Wikipedia war narratives…" authors: ["Jannik Strötgen", "Michael Gertz"] year: 2012 venue: "Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)" source: "https://aclanthology.org/L12-1219/" licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)" licence_url: "https://creativecommons.org/licenses/by-nc/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/strotgen-2012-temporal-tagging-different-domains-challenges/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/strotgen-2012-temporal-tagging-different-domains-challenges/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2012](https://md.research.elamur.ai/free/en/papers/2012/) / Temporal Tagging on Different Domains: Challenges, Strategies, and Gold Standards # Temporal Tagging on Different Domains: Challenges, Strategies, and Gold Standards **Authors:** [Jannik Strötgen](https://md.research.elamur.ai/free/en/authors/jannik-strotgen/), [Michael Gertz](https://md.research.elamur.ai/free/en/authors/michael-gertz/) · **Year:** 2012 · **Venue:** Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12) · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/) **Links:** [Publisher's page](https://aclanthology.org/L12-1219/) · [PDF](https://aclanthology.org/L12-1219.pdf) · [OpenAlex](https://openalex.org/W2250852500) **Citations:** 40 citations (openalex-snapshot · 2026-10-14) · [80](#cited-by) in our library, of the 140673 papers whose bibliographies we read ## Summary Most temporal taggers and annotated corpora come from news, and earlier work showed that moving to other text types, such as Wikipedia war narratives, lowers both extraction and normalization quality. Strötgen and Gertz ask how temporal expressions differ across four domains (news, narrative, colloquial SMS and scientific biomedical text) and whether telling the tagger the domain helps. They built two new TIMEX2-annotated gold standards. Time4SMS is drawn from the NUS SMS corpus, and Time4SCI is biomedical text. They also analysed the domain-specific difficulties: reference-time resolution, informal spelling, and relative chains with no anchor date. They added matching strategies and resources to HeidelTime, which now has news, narrative, colloquial and scientific settings. In a cross-domain evaluation, the matching domain setting gave the best results on every corpus. The largest gains were in normalization. The work argues for domain-aware temporal tagging and releases the corpora and tagger publicly. ## Key points - The colloquial corpus is based on the NUS SMS corpus, June 2011 version, with 28,268 messages. The 2004 portion was excluded because it has no sending times, which are needed to resolve relative and underspecified expressions. Anonymization placeholders for digits and times were replaced with standard values in the original format. - The scientific domain uses a "time point zero" (TPZ) normalization. Relative expressions in a chain are accumulated against a document-internal origin. For example, "two days later" followed by "one day later" yields TPZ+0000-00-02 and then TPZ+0000-00-03. - The new HeidelTime version takes both a language and a domain setting. The four configurations are english/news, english/narrative, english-coll/colloquial and english-sci/scientific, with additional vocabulary and rule resources for the colloquial and scientific settings. - Evaluation used precision, recall and F-score under five settings: lenient, strict, value, len+val and str+val. Value-normalization F-scores show the effect of choosing the right setting: - TimeBank: 86.2 with the news setting vs 67.5 with narrative. - WikiWars: 89.8 with narrative vs 64.9 with news. - Time4SCI: 88.7 with scientific vs 74.4 with news. - Time4SMS: lenient extraction F rose from 91.7 to 95.1 with the colloquial setting. - Time4SMS scores well under every setting because of its many document creation times (1000). The colloquial setting still performs best there. ## Interesting ideas - **Domain as an explicit input.** The user declares the domain at run time instead of relying on one universal rule set. A single tool can then switch reference-time and normalization logic, which would suit Ukrainian NLP pipelines that handle news, forums and legal text side by side. - **Relative timelines without an absolute anchor.** Normalizing to an offset from a document-internal zero point keeps chronological information that would otherwise collapse into an unknown value. The same approach could support procedural texts, trade logs or route descriptions that lack calendar dates. - **Explicit criteria for choosing source data.** Before picking a base corpus, the authors listed requirements: free licence, English, large size, and available creation timestamps. Writing such criteria down makes corpus choices reproducible and defensible. - **Restoring anonymized fields with format-preserving values.** Placeholders were swapped for standard values in the original format, so privacy-scrubbed data remained usable for a task that depends on numbers and times. This could help with other anonymized datasets. - **Running every model configuration on every domain.** Evaluating all settings against all corpora at several strictness levels separates extraction errors from normalization errors. It also shows exactly where domain adaptation pays off. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* Fields: [Linguistics & languages](https://md.research.elamur.ai/free/en/shelves/linguistics/) Topics: [Part-of-speech tagging](https://md.research.elamur.ai/free/en/topics/part-of-speech-tagging/) ## Cited by Cited by 80 papers of our library cite it (of the 140673 whose bibliographies we read); 3 of them have a page here, 77 not yet. - [WordNet under Scrutiny: Dictionary Examples in the Era of Large Language Models](https://md.research.elamur.ai/free/en/papers/lex-2024-almeman-wordnet-examples-llm/) - [MindSET: Advancing Mental Health Benchmarking through Large-Scale Social Media Data](https://md.research.elamur.ai/free/en/papers/mankarious-2026-mindset-advancing-mental-health-benchmarking/) - [A Parallel Cross-Lingual Benchmark for Multimodal Idiomaticity Understanding](https://md.research.elamur.ai/free/en/papers/selamet-2026-parallel-cross-lingual-benchmark-multimodal/) ## Cite The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{strotgen2012temporal, title = {{Temporal Tagging on Different Domains: Challenges, Strategies, and Gold Standards}}, author = {Strötgen, Jannik and Gertz, Michael}, booktitle = {Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)}, year = {2012}, doi = {10.63317/4q224cfvphx9}, url = {https://doi.org/10.63317/4q224cfvphx9} } ``` - [← previous Syntactic annotation of spontaneous speech: application to call-center conversation data](https://md.research.elamur.ai/free/en/papers/bazillon-2012-syntactic-annotation-spontaneous-speech-application/) - [next → Text-level Discourse Parsing with Rich Linguistic Features](https://md.research.elamur.ai/free/en/papers/feng-2012-text-level-discourse-parsing-rich/) - [↑ back to papers of 2012](https://md.research.elamur.ai/free/en/papers/2012/) ## References 3 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Tao Chen and Min-Yen Kan. 2011. Creating a Live, Public Short Message Service Corpus: The NUS SMS Corpus. Technical report, National University of Singapore. [find](https://scholar.google.com/scholar?q=Tao+Chen+and+Min%2DYen+Kan%2E+2011%2E+Creating+a+Live%2C+Public+Short+Message+Service+Corpus%3A+The+NUS+SMS+Corpus%2E+Technical+report%2C+National+University+of+Singapore%2E) 2. Lisa Ferro, Laurie Gerber, Inderjeet Mani, Beth Sundheim, and George Wilson. 2005. [find](https://scholar.google.com/scholar?q=Lisa+Ferro%2C+Laurie+Gerber%2C+Inderjeet+Mani%2C+Beth+Sundheim%2C+and+George+Wilson%2E+2005%2E) 3. TIDES 2005 Standard for the [find](https://scholar.google.com/scholar?q=TIDES+2005+Standard+for+the)