2019
DOI: 10.21605/cukurovaummfd.609119
|View full text |Cite
|
Sign up to set email alerts
|

Investigation of Word Similarities with Word Embedding Methods

Abstract: Günümüzde büyük veri alanında meydana gelen gelişmelerle birlikte günlük işlenebilir durumda olan veri miktarı oldukça büyük boyutlara ulaşmıştır. Bu verilerin çok büyük bir kısmının metin (text) verilerinden oluşması, metin işleme alanında yapılan çalışmaları oldukça önemli ve popüler bir hale getirmiştir. Ancak bu alanda yapılan çalışmalar incelendiğinde başta İngilizce olmak üzere birçok dünya diline yönelik çeşitli çalışmalar yapılırken, Türkçeye özgü yapılan çalışmaların istenilen sayıda olmadığı görülmüş… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

1
4
0

Year Published

2020
2020
2024
2024

Publication Types

Select...
4
1

Relationship

1
4

Authors

Journals

citations
Cited by 11 publications
(5 citation statements)
references
References 13 publications
1
4
0
Order By: Relevance
“…Among the static embedding models, Word2Vec and FastText outperform GloVe in the context of the Turkish language. This finding aligns with the results obtained by (Aydogan & Karcı, 2019).…”
Section: Intrinsic Evaluationsupporting
confidence: 93%
See 1 more Smart Citation
“…Among the static embedding models, Word2Vec and FastText outperform GloVe in the context of the Turkish language. This finding aligns with the results obtained by (Aydogan & Karcı, 2019).…”
Section: Intrinsic Evaluationsupporting
confidence: 93%
“…They also demonstrated that manipulating words' surface forms can genuinely enhance the efficacy of text representation in a morphologically rich language. (Aydogan & Karcı, 2019) conducted a comparison between Word2Vec and GloVe models, evaluating their performance on analogy tasks and the training time. Within the scope of the study, the two fundamental algorithms of the Word2Vec method, CBOW and Skip-gram, have also been analyzed.…”
Section: Related Workmentioning
confidence: 99%
“…But in SkipGram, the approach is to give a word to a feed-forward neural network and predict the words around it. With the Word-2Vec method [12] developed by Google researchers, it become possible to represent any word on a corpus of billions of words with a 300-dimensional vector space. The Glove method (Global Vectors for Word Representation) [4] is another type of word vector generation method from huge corpora in an unsupervised manner.…”
Section: Introductionmentioning
confidence: 99%
“…While many studies in English examine semantic similarity tasks using word vectors, a limited number of studies can be found for Turkish [12,13,14]. Aydogan and Karci (2019) [12] used the Beautiful Soup library to create a large corpus of 60GB Turkish texts, containing 10.5 billion words, and created word representation vectors by training this corpus with both Word2Vec (CBOW and SkipGram) and Glove methods.…”
Section: Introductionmentioning
confidence: 99%
See 1 more Smart Citation