| Título |
The Winning Approach to Cross-Genre Gender Identification in Russian at RUSProfiling 2017 |
| Tipo |
Congreso |
| Sub-tipo |
Memoria |
| Descripción |
2017 Working Notes of Forum for Information Retrieval Evaluation, FIRE 2017 |
| Resumen |
WepresenttheCICsystemssubmittedtothe2017PANsharedtask on Cross-Genre Gender Identification in Russian texts (RUSProfiling). We submitted five systems. One of them was based on a statisticalapproachusingonlylexicalfeatures,andotherfouron machine-learningtechniquesusingsomecombinationsofgenderspecificRussiangrammaticalfeatures,wordandcharactern-grams, and suffix n-grams. Our systems achieved the highest weighted accuracyacrossallthetestdatasets,occupyingthefirstfourplaces intheranking.
|
| Observaciones |
http://ceur-ws.org/Vol-2036/T1-5.pdf,
CEUR Workshop Proceedings, v. 2036 |
| Lugar |
Bangalore |
| País |
India |
| No. de páginas |
20-24 |
| Vol. / Cap. |
|
| Inicio |
2017-12-08 |
| Fin |
2017-12-10 |
| ISBN/ISSN |
|
