Matches in SemOpenAlex for { <https://semopenalex.org/work/W3037917809> ?p ?o ?g. }
Showing items 1 to 47 of
47
with 100 items per page.
- W3037917809 abstract "Nous nous interessons dans cet article a l'apprentissage automatique d'un etiqueteur mopho-syntaxique pour les tweets en anglais. Nous proposons tout d'abord un jeu d'etiquettes reduit avec 17 etiquettes differentes, qui permet d'obtenir de meilleures performances en exactitude par rapport au jeu d'etiquettes traditionnel qui contient 45 etiquettes. Comme nous disposons de peu de tweets etiquetes, nous essayons ensuite de compenser ce handicap en ajoutant dans l'ensemble d'apprentissage des donnees issues de textes bien formes. Les modeles mixtes obtenus permettent d'ameliorer les resultats par rapport aux modeles appris avec un seul corpus, qu'il soit issu de Twitter ou de textes journalistiques. Abstract. Part-of-speech Tagging for Tweets with CRFs We are insterested in this paper in training a part-of-speach tagger for tweets in English. We first propose a reduced tagset with 17 different tags, which allows better results in accuracy than traditional tagsets which contain 45 tags. Since we have few annoted tweets, we then try and overcome this difficulty by adding data from other more standard texts into the training set. The obtained models reach better results compared to models trained with only one corpus, whether coming of Twitter or of journalistic texts. Mots-cles : tweets, CRF, etiquettage morpho-syntaxique." @default.
- W3037917809 created "2020-07-02" @default.
- W3037917809 creator A5050707731 @default.
- W3037917809 creator A5063627915 @default.
- W3037917809 creator A5076187752 @default.
- W3037917809 creator A5091335565 @default.
- W3037917809 date "2015-06-22" @default.
- W3037917809 modified "2023-10-02" @default.
- W3037917809 title "Étiquetage morpho-syntaxique de tweets avec des CRF" @default.
- W3037917809 hasPublicationYear "2015" @default.
- W3037917809 type Work @default.
- W3037917809 sameAs 3037917809 @default.
- W3037917809 citedByCount "0" @default.
- W3037917809 crossrefType "proceedings-article" @default.
- W3037917809 hasAuthorship W3037917809A5050707731 @default.
- W3037917809 hasAuthorship W3037917809A5063627915 @default.
- W3037917809 hasAuthorship W3037917809A5076187752 @default.
- W3037917809 hasAuthorship W3037917809A5091335565 @default.
- W3037917809 hasBestOaLocation W30379178091 @default.
- W3037917809 hasConcept C154945302 @default.
- W3037917809 hasConcept C2777261249 @default.
- W3037917809 hasConcept C41008148 @default.
- W3037917809 hasConcept C59822182 @default.
- W3037917809 hasConcept C86803240 @default.
- W3037917809 hasConceptScore W3037917809C154945302 @default.
- W3037917809 hasConceptScore W3037917809C2777261249 @default.
- W3037917809 hasConceptScore W3037917809C41008148 @default.
- W3037917809 hasConceptScore W3037917809C59822182 @default.
- W3037917809 hasConceptScore W3037917809C86803240 @default.
- W3037917809 hasLocation W30379178091 @default.
- W3037917809 hasLocation W30379178092 @default.
- W3037917809 hasOpenAccess W3037917809 @default.
- W3037917809 hasPrimaryLocation W30379178091 @default.
- W3037917809 hasRelatedWork W2187479637 @default.
- W3037917809 hasRelatedWork W2211302168 @default.
- W3037917809 hasRelatedWork W2478840300 @default.
- W3037917809 hasRelatedWork W2569931262 @default.
- W3037917809 hasRelatedWork W2748952813 @default.
- W3037917809 hasRelatedWork W2899084033 @default.
- W3037917809 hasRelatedWork W4211036162 @default.
- W3037917809 hasRelatedWork W4245825737 @default.
- W3037917809 hasRelatedWork W4252623201 @default.
- W3037917809 hasRelatedWork W4300887996 @default.
- W3037917809 isParatext "false" @default.
- W3037917809 isRetracted "false" @default.
- W3037917809 magId "3037917809" @default.
- W3037917809 workType "article" @default.