Tracking information flow (IFLOW) is crucial to understanding the evolution of news stories. We present analysis and experiments for IFLOW between company announcements and newswire. Error analysis shows that many FPs are annotation errors and many FNs are due to coarse-grained document-level modelling. Experiments show that document meta-data features (e.g., category, length, timing) improve f-scores relative to upper bound by 23%.
|Title of host publication||Proceedings of the Main Conference on Human Language Technologies|
|Subtitle of host publication||The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL HLT 2010)|
|Place of Publication||Stroudsburg, PA|
|Publisher||Association for Computational Linguistics|
|Number of pages||2|
|Publication status||Published - 2010|
|Event||North American Association for Computational Linguistics - Los Angeles, CA, United States|
Duration: 2 Jun 2010 → 6 Jun 2010
|Conference||North American Association for Computational Linguistics|
|City||Los Angeles, CA|
|Period||2/06/10 → 6/06/10|
Radford, W., Hachey, B., Curran, J. R., & Milosavljevic, M. (2010). Tracking Information Flow between Primary and Secondary News Sources. In Proceedings of the Main Conference on Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL HLT 2010) (pp. 29-30). Stroudsburg, PA: Association for Computational Linguistics.