A comprehensive analysis of static word embeddings for Turkish

Open Webpage

Karahan Saritas, Cahid Arda Öz, Tunga Güngör

Published: 01 Jan 2024, Last Modified: 04 Nov 2024Expert Syst. Appl. 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•A comprehensive analysis of static word embedding models for the Turkish language.•Intrinsic and extrinsic performances of the models on several tasks are presented.•BERT with X2Static method outperformed all models in intrinsic and extrinsic tasks.•FastText excels in conjugation tasks, while word2vec shines in analogy tasks.