Year: 2024
Type: article
Abstract: Spell checking is the task of detecting and correcting spelling errors in text and is one of the most sought-after processes in NLP. There are many open-source toolkits for checking and correcting errors in the text. To test how effective these tools are, in this article I have presented an evaluation of three types of tools as NeuSpell, SymSpell and Hunspell. SymSpell showed a high speed of 2480, this is an indicator of how fast it works than others. And NeuSpell achieved the lowest error rate of 0.80%. The results show the disadvantages and advantages of all algorithms, and that there is still room for improvement.
Author: Assylay Tolegenova
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4407232086