: Vocabulaire du TAL: blend

computational linguistics > morphology > blend

blend

On a first approximation, blends (Ger. Kontamination, Wortverschmelzung, Fr. mots-valises, amalgames, It. parole macedonia, incroci) are lexemes formed by means of fusing two already existing lexemes into a new one, where the stems of the initial lexemes have often been shortened. English smog ← smoke × fog is an example of a blend. (Fradin, Bernard. "21. Blending". Volume 1 Word-Formation: An International Handbook of the Languages of Europe, edited by Peter O. Müller, Ingeborg Ohnheiser, Susan Olsen and Franz Rainer, Berlin, München, Boston: De Gruyter Mouton, 2015, pp. 386-413. https://doi.org/10.1515/9783110246254-023)

2. Blends or portmanteaus are formed by merging parts of existing words. (Kulkarni & Wang, 2018)
In all the results reported character level accuracy indicates the generalized accuracy over number of characters that were predicted correctly in the blend word. (Das & Ghosh, 2017)
The dominance of the Retro model and the SIG nonneur baseline could be (at least partially) caused by the fact that test-neologisms contains high percentage (37%) of compounds blends or words derived by prefixing whose ending segment is an existing word present in MorfFlex (and thus possibly present in the training data). (Sourada, Straková & Rosa, 2024)
While portmanteaus (word blends) word combinations and other formation mechanisms do exist in Hebrew most words are created by combination of root and pattern. (Mizrahi, Yardeni Seelig & Shahaf, 2020)
Word level accuracy indicates the generalized accuracy over the number of correct blend word predictions over all the source word pair instances. (Das & Ghosh, 2017)

http://data.loterre.fr/ark:/67375/8LP-M9V1BD92-N

RDF/XML TURTLE JSON-LD Date de création 26/06/2024, dernière modification le 28/06/2024

Loterre