Context Based Automatic Spelling Correction for Turkish

Bölücü, Necva; Can, Burcu

doi:10.1109/ebbt.2019.8742067

Cited by 3 publications

(3 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the previous Turkish spelling correction studies (Bolucu & Can, 2019;Buyuk, 2020), it is observed that incorporating context information significantly improves the correction performance. We also employ a contextual spelling correction as in Buyuk (2020).…”

Section: Context-dependent Spelling Correctionmentioning

confidence: 94%

“…Every walk starts from a misspelled word and ends at a correct word with an associated cost. Similar to ETFSR, Zemberek produces multiple correction candidates for each misspelling (Bolucu & Can, 2019). In Bolucu & Can (2019), Zemberek's correction candidates are ranked using context information.…”

mentioning

confidence: 99%

“…Similar to ETFSR, Zemberek produces multiple correction candidates for each misspelling (Bolucu & Can, 2019). In Bolucu & Can (2019), Zemberek's correction candidates are ranked using context information. In the method, hidden Markov model (HMM) parameters (initial distribution, transition and emission probabilities) are trained using the text corpus in Sak et al (2008).…”

mentioning

confidence: 99%

See 2 more Smart Citations

Learning from mistakes: Improving spelling correction performance with automatic generation of realistic misspellings

Büyük

Arslan

2021

Expert Systems

View full text Add to dashboard Cite

Sequence to sequence models (seq2seq) require a large amount of labelled training data to learn the mapping between the input and output. A large set of misspelled words together with their corrections is needed to train a seq2seq spelling correction system. Low-resource languages such as Turkish usually lack such large annotated datasets. Although misspelling-reference pairs can be synthesized with a random procedure, the generated dataset may not well match to genuine human-made misspellings. This might degrade the performance in realistic test scenarios. In this paper, we propose a novel procedure to automatically introduce human-like misspellings to legitimate words in Turkish language. Generated human-like misspellings are used to improve the performance of a seq2seq spelling correction system. The proposed system consists of two separate models; a misspelling generator and a spelling corrector.The generator is trained using a relatively small number of human-made misspellings and their manual corrections. Reference words and their misspellings are used as inputs and outputs of the generator, respectively. As a result, it is trained to add realistic spelling errors to the valid words. Training data of the spelling corrector is augmented by the generator's human-like misspellings. In the experiments, we observe that the data augmentation significantly improves the spelling correction performance. Our proposed method yields 5% absolute improvement over the state-ofthe-art Turkish spelling correction systems in a test set which contains human-made misspellings from Twitter messages.

show abstract

Section: Context-dependent Spelling Correctionmentioning

confidence: 94%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Learning from mistakes: Improving spelling correction performance with automatic generation of realistic misspellings

Büyük

Arslan

2021

Expert Systems

View full text Add to dashboard Cite

show abstract

#Turki$hTweets: A Benchmark Dataset for Turkish Text Correction

Koksal¹,

Bozal²,

Yürekli³

et al. 2020

Findings of the Association for Computational Linguistics: EMNLP 2020

View full text Add to dashboard Cite

Turki$hTweets is a benchmark dataset for the task of correcting the user misspellings, with the purpose of introducing the first public Turkish dataset in this area. #Turki$hTweets provides correct/incorrect word annotations with a detailed misspelling category formulation based on the real user data. We evaluated four state-of-the-art approaches on our dataset to present a preliminary analysis for the sake of reproducibility. The annotated dataset is publicly available at https://github.com/ atubakoksal/annotated_tweets.

show abstract

TurkishLex:Development of a Context-Aware Spell Checker for Detecting and Correcting Spelling Errors in Turkish Texts

Savci,

Daş

2024

SAUCIS

View full text Add to dashboard Cite

In Turkish, correct spelling correction is crucial for effective communication and preserving the integrity of written text. The challenge lies in the complexity of Turkish morphology and spelling, which can lead to frequent and diverse spelling errors. This study presents a spelling checker adapted for Turkish by creating a new Turkish dataset. The proposed spelling checker model effectively captures both minor and major textual changes and can detect the error. Our findings show that the proposed spelling checker system provides high accuracy and reliability with 98.21% accuracy performance with the Symspell module in correcting Turkish texts. This study provides valuable information about the strengths and weaknesses of existing spelling checkers and contributes to the improvement of spelling correction tools for Turkish.

show abstract

Context Based Automatic Spelling Correction for Turkish

Cited by 3 publications

References 9 publications

Learning from mistakes: Improving spelling correction performance with automatic generation of realistic misspellings

Learning from mistakes: Improving spelling correction performance with automatic generation of realistic misspellings

#Turki$hTweets: A Benchmark Dataset for Turkish Text Correction

TurkishLex:Development of a Context-Aware Spell Checker for Detecting and Correcting Spelling Errors in Turkish Texts

Contact Info

Product

Resources

About