Difference between revisions of "POS Tagging (State of the art)"
Jump to navigation
Jump to search
Line 1: | Line 1: | ||
+ | |||
+ | == "Standard" measure: == | ||
+ | * Per token accuracy | ||
+ | |||
+ | == "Standard" datasets: == | ||
+ | * Training: sections 0-18 of WSJ | ||
+ | * Testing: sections 22-24 of WSJ | ||
+ | |||
+ | |||
{{StateOfTheArtTable}} | {{StateOfTheArtTable}} | ||
− | | SVMTool || SVM Based tagger and tagger generator || Jesús Giménez and Lluís Márquez. SVMTool: A general POS tagger generator based on Support Vector Machines[http://www.lsi.upc.es/~nlp/SVMTool/lrec2004-gm.pdf] || [http://www.lsi.upc.es/~nlp/SVMTool/|http://www.lsi.upc.es/~nlp/SVMTool/ || 97.16% | + | | SVMTool || SVM Based tagger and tagger generator || Jesús Giménez and Lluís Márquez. SVMTool: A general POS tagger generator based on Support Vector Machines[http://www.lsi.upc.es/~nlp/SVMTool/lrec2004-gm.pdf] || [http://www.lsi.upc.es/~nlp/SVMTool/|http://www.lsi.upc.es/~nlp/SVMTool/ || 97.16% || |
+ | |- | ||
+ | |||
+ | | --- || Learning with Cyclic Dependency Network || Kristina Toutanova, Dan Klein, Christopher D. Manning, and Yoram Singer. Feature-Rich Part-of-Speech Tagging with a Cyclic Dependency Network [http://nlp.stanford.edu/kristina/papers/tagging.pdf] || No || 97.24% || | ||
|- | |- | ||
|} | |} |
Revision as of 10:56, 16 June 2007
"Standard" measure:
- Per token accuracy
"Standard" datasets:
- Training: sections 0-18 of WSJ
- Testing: sections 22-24 of WSJ
System Name | Short Description | Main Publications | Software (if available) | Results | Comments (i.e. extra resources used, train/test times, ...) |
---|---|---|---|---|---|
SVMTool | SVM Based tagger and tagger generator | Jesús Giménez and Lluís Márquez. SVMTool: A general POS tagger generator based on Support Vector Machines[1] | http://www.lsi.upc.es/~nlp/SVMTool/ | 97.16% | |
--- | Learning with Cyclic Dependency Network | Kristina Toutanova, Dan Klein, Christopher D. Manning, and Yoram Singer. Feature-Rich Part-of-Speech Tagging with a Cyclic Dependency Network [2] | No | 97.24% |