Difference between revisions of "POS Induction (State of the art)"
Jump to navigation
Jump to search
(Berg-Kirkpatrick et al. (2010)) |
(Alexander Clark. 2003) |
||
| Line 1: | Line 1: | ||
| + | ==Evaluation== | ||
| + | '''Many-to-1:''' Mapping every induced label to a gold standard tag greedily. Use the mapping to compute tag accuracy on the Wall Street Journal part of the Penn TreeBank. | ||
| + | |||
| + | ==Results== | ||
| + | |||
{| border="1" cellpadding="5" cellspacing="1" width="100%" | {| border="1" cellpadding="5" cellspacing="1" width="100%" | ||
|- | |- | ||
| Line 7: | Line 12: | ||
! Many-to-1 | ! Many-to-1 | ||
|- | |- | ||
| − | | | + | | Brown+proto |
| MRF initialized with Brown prototypes | | MRF initialized with Brown prototypes | ||
| Christodoulopoulos, Goldwater and Steedman (2010) | | Christodoulopoulos, Goldwater and Steedman (2010) | ||
| Line 18: | Line 23: | ||
| | | | ||
| 75.5% | | 75.5% | ||
| + | |- | ||
| + | | Clark DMF | ||
| + | | Distributional clustering + morphology + frequency | ||
| + | | Clark (2003) | ||
| + | | [http://www.cs.rhul.ac.uk/home/alexc/pos2.tar.gz alexc] | ||
| + | | 71.2%* | ||
|- | |- | ||
|} | |} | ||
| + | <nowiki>*</nowiki> according to Christodoulopoulos, Goldwater and Steedman (2010) | ||
== References == | == References == | ||
| Line 27: | Line 39: | ||
* [http://www.aclweb.org/anthology/N/N10/N10-1083.pdf Taylor Berg-Kirkpatrick, Alexandre Bouchard-Cote, John DeNero, and Dan Klein. 2010. Painless Unsupervised Learning with Features. NAACL 2010.] | * [http://www.aclweb.org/anthology/N/N10/N10-1083.pdf Taylor Berg-Kirkpatrick, Alexandre Bouchard-Cote, John DeNero, and Dan Klein. 2010. Painless Unsupervised Learning with Features. NAACL 2010.] | ||
| + | |||
| + | * [http://www.aclweb.org/anthology/E/E03/E03-1009.pdf Alexander Clark. 2003. Combining distributional and morphological information for part of speech induction. In Proceedings of EACL 2003, pages 59–66, Morristown, NJ, USA.] | ||
== See also == | == See also == | ||
Revision as of 15:43, 27 January 2011
Evaluation
Many-to-1: Mapping every induced label to a gold standard tag greedily. Use the mapping to compute tag accuracy on the Wall Street Journal part of the Penn TreeBank.
Results
| System name | Short description | Main publications | Software | Many-to-1 |
|---|---|---|---|---|
| Brown+proto | MRF initialized with Brown prototypes | Christodoulopoulos, Goldwater and Steedman (2010) | 76.1% | |
| Logistic regression with features and LBFGS | Berg-Kirkpatrick et al. (2010) | 75.5% | ||
| Clark DMF | Distributional clustering + morphology + frequency | Clark (2003) | alexc | 71.2%* |
* according to Christodoulopoulos, Goldwater and Steedman (2010)