Tagger: Difference between revisions

From Glottopedia
Jump to navigation Jump to search
New page: ==Definition== Ein Tagger versieht jedes Wort in einem Korpus mit einem Tag (Etikett), welches eine part of speech oder eine andere lexikalische Kategorie Kategorie bezeichnet. Die...
 
NBlöcher (talk | contribs)
Marked as {{ref}}
 
(11 intermediate revisions by 2 users not shown)
Line 1: Line 1:
==Definition==
==Definition==
Ein Tagger versieht jedes Wort in einem [[Korpus]] mit einem Tag (Etikett), welches eine part of speech oder eine andere [[lexikalische Kategorie]] Kategorie bezeichnet. Die Tags stammen aus einem Tagset.  
A '''tagger''' is a device which assigns symbolic labels (''tags'') to linguistics units. The labels are taken from a predefined set of symbols (the so-called ''tag-set'').


Man unterscheidet zwischen regelbasierten und stochastischen Taggern. Stochastische (oder statistische) Tagger arbeiten mit Übergangswahrscheinlichkeiten von Wortfolgen, während regelbasierte Tagger (oder Brill-Tagger) Regeln zur Wortfolge aufstellen, die mit weiteren Regeln inkrementell verbessert werden. Beide Tagging-Algorithmen lernen die erforderlichen Informationen aus einem Trainingskorpus.
==Comments==
Tagger erreichen je nach Textsorte eine Korrektheit von 90-97%.
In most cases, a tagger assigns tags representing morpho-syntactic information to single word-forms or tokens. But there are also taggers which have been designed to identify semantic roles of noun phrases or prepositional phrases ([[sense tagging]]), and sometimes identifying the structure of a text is considered as a kind of tagging ([[discourse structure tagging]]).


==Herkunft==
Conceptually, tagging can be considered as a three step process: (i). identification of the relevant units (ii). assigning all possible labels (e.g. by lexical look-up, applying heuristics, etc.) (iii). disambiguation.
engl. ''tag'' - markieren, mit einem Anhängeretikett versehen
 
It is common practice to distinguish between rule-based and stochastic taggers, though some taggers combine rules and stochastic information.
In general, state-of-the-art taggers achieve a precision of at least 95% for morpho-syntactic tagging.
 
==Subtypes==
* [[Brill tagger]]
* [[HMM tagger]]
* [[Memory-based tagger]]
* [[Tree tagger]]
 
==Other Languages==
 
* German [[Tagger (de)]]
 
{{dc}}{{ref}}
[[Category:Computational Linguistics]]

Latest revision as of 16:30, 27 July 2014

Definition

A tagger is a device which assigns symbolic labels (tags) to linguistics units. The labels are taken from a predefined set of symbols (the so-called tag-set).

Comments

In most cases, a tagger assigns tags representing morpho-syntactic information to single word-forms or tokens. But there are also taggers which have been designed to identify semantic roles of noun phrases or prepositional phrases (sense tagging), and sometimes identifying the structure of a text is considered as a kind of tagging (discourse structure tagging).

Conceptually, tagging can be considered as a three step process: (i). identification of the relevant units (ii). assigning all possible labels (e.g. by lexical look-up, applying heuristics, etc.) (iii). disambiguation.

It is common practice to distinguish between rule-based and stochastic taggers, though some taggers combine rules and stochastic information.

In general, state-of-the-art taggers achieve a precision of at least 95% for morpho-syntactic tagging.

Subtypes

Other Languages

REF This article has no reference(s) or source(s).
Please remove this block only when the problem is solved.