Automatické rozpoznávání substantivního dativu při morfologické disambiguaci českých textů

Title:

Automatické rozpoznávání substantivního dativu při morfologické disambiguaci českých textů
Automatic identification of the nominal dative in a morphological disambiguation of Czech texts

Creator:

Petkevič, Vladimír

Identifier:

https://cdk.lib.cas.cz/client/handle/uuid:d79fd3f7-775c-48b0-82db-4ee550af55cf
uuid:d79fd3f7-775c-48b0-82db-4ee550af55cf

Subject:

non-prepositional nominal dative, corpus of Czech texts, morphological disambiguation, positive and negative rules, nepředložkový dativ substantiva, korpus českých textů, morfologická disambiguace, and pozitivní a negativní pravidla

Type:

model:article and TEXT

Format:

bez média and svazek

Description:

The paper deals with the identification of the dative/non-dative interpretation of a non-prepositional noun (the head noun in a prepositional group) in a rule-based automatic morphological disambiguation of Czech sentences. Based on the valency considerations including syntactic functions of the complements of verbs (these functions being an object or an adverbial) and resulting in the lists of verbs having / not having object dative valency, negative and positive non-heuristic (safe) as well as heuristic rules are presented. The negative rules assign a noun a non-dative interpretation using a list of verbs that cannot tolerate a dative noun in their close vicinity, whereas the positive ones specify a given noun as being in the dative case on the basis of the verbs having an object dative valency, an object being either the only object of the verb, or its second indirect object. The paper deals only with the non-prepositional nouns which are much more difficult to disambiguate than the prepositional ones.

Language:

Czech

Rights:

http://creativecommons.org/publicdomain/mark/1.0/
policy:public

Coverage:

67-76

Source:

Korpus - gramatika - axiologie: časopis pro korpusový výzkum a hodnocení jazyka | 2013 | Number:7

Harvested from:

CDK

Metadata only:

false