Abstract | ||
---|---|---|
Orthographic variation can be a serious problem for many nat- ural language-processing applica- tions. Japanese in particular con- tains orthographic variation, be- cause the large quantity of translit- eration from other languages causes many possible spelling variations. To manage this problem, this pa- per proposes a support vector ma- chine (SVM)-based classifier that can determine whether two terms are equivalent. We automatically collected both positive examples (sets of equivalent term pairs) and negative examples (sets of inequiv- alent term pairs). Experimental re- sults yielded high levels of accuracy (87.8%), demonstrating the feasibil- ity of the proposed approach. |
Year | Venue | DocType |
---|---|---|
2007 | TMI | Conference |
Citations | PageRank | References |
1 | 0.35 | 6 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Eiji Aramaki | 1 | 371 | 45.89 |
Takeshi Imai | 2 | 20 | 3.73 |
Kengo Miyo | 3 | 24 | 5.65 |
Kazuhiko Ohe | 4 | 115 | 15.91 |