2004
DOI: 10.1016/j.tcs.2003.10.007
|View full text |Cite
|
Sign up to set email alerts
|

Finite-state transducer cascades to extract named entities in texts

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
31
0
5

Year Published

2006
2006
2019
2019

Publication Types

Select...
4
3
2

Relationship

0
9

Authors

Journals

citations
Cited by 63 publications
(36 citation statements)
references
References 11 publications
0
31
0
5
Order By: Relevance
“…However, many knowledge-based approaches have been proposed in the field of NER, one of the earliest being based on heuristics and handcrafted rules [29]. Many different methods are used, such as cascades of finite-state transducers that produce tree-like representations [8,28]. Because regular languages and relations can be encoded as finite automata, they can be more easily manipulated than more complex languages; cascades of transducers have therefore turned out to be very useful for many approaches based on domain-specific corpus analysis and rules are described in a readable way and are easy to modify and maintain.…”
Section: Named Entity Recognition (Ner)mentioning
confidence: 99%
See 1 more Smart Citation
“…However, many knowledge-based approaches have been proposed in the field of NER, one of the earliest being based on heuristics and handcrafted rules [29]. Many different methods are used, such as cascades of finite-state transducers that produce tree-like representations [8,28]. Because regular languages and relations can be encoded as finite automata, they can be more easily manipulated than more complex languages; cascades of transducers have therefore turned out to be very useful for many approaches based on domain-specific corpus analysis and rules are described in a readable way and are easy to modify and maintain.…”
Section: Named Entity Recognition (Ner)mentioning
confidence: 99%
“…In this body of work, a significant part is devoted to the study of English, but French is also considered [8,27], as well as some other languages. The impact of the literary genre (e.g., narrative, memoir or journalism) and domain (e.g., supply of raw materials, market or economic intelligence, or politics) is a problem that has been more recently addressed in the NER literature.…”
Section: Named Entity Recognition (Ner)mentioning
confidence: 99%
“…Repérage et caractérisation manuel des Entités Nommées, et, plus largement, des éléments principaux d'une chaîne anaphorique (pronoms et Groupes Nominaux) par un annotateur. Les entités nommées ont été automatiquement identifiées en utilisant la ressource CasEN (tln.li.univtours.fr/Tln_CasEN.html) avec le logiciel CasSys, disponible sur la plateforme Unitext (Friburger et Maurel, 2004 ;Maurel et alii, 2011), puis corrigées (uniquement pour le corpus ESLO) par un expert humain selon la convention ESTER2 (Galliano et alii, 2005). Les autres GN, dont les pronoms, ont été identifiés semi-automatiquement avant la tâche d'annotation des anaphores ; 4.…”
Section: Procédure D'annotationunclassified
“…Le texte français subit alors une étape fondamentale : le repérage et l'étiquetage des noms propres qu'il contient. Nous avons réalisé l'extraction des noms propres dans le texte en version originale grâce à l'outil CasSys (Friburger, 2004), disponible sur la plateforme de traitement linguistique Unitex 7 (Paumier, 2006), et aux transducteurs regroupés sous le nom de CasEN 8 , développés pour la campagne d'évaluation des Systèmes de Transcription Enrichie d'Emissions Radiophoniques (ESTER) 9 . Après une phase de pré-analyse (division du texte en phrases, étiquetage avec les dictionnaires, etc.…”
Section: Présentation Des Outilsunclassified