• Yusufjanov Ulugbek Javlon Ugli

  Author

• Shermatov Abdukodir Obidjon Ugli

  Author

The emergence of AI and NLP technologies is a major driving force behind the transformation of the field of English language acquisition, with automated writing assistance being introduced in order to solve a number of challenges associated with learning English. Nevertheless, English language students still have problems with identifying and correcting grammatical errors due to their limited language exposure, the influence of the native language, and a lack of continuous feedback. This study aims to assess the efficiency of an NLP-based Automatic Grammatical Error Correction (AGEC) approach intended to support English writing proficiency by means of context-based error detection, correction generation, and provision of pedagogically relevant feedback. The AGE Correction Framework suggested in the study involves modern NLP techniques such as transformers and Large Language Models (LLMs). The analysis of learner-generated text material, as well as correction accuracy and error-handling ability of the framework, is performed with the use of learner writing datasets comprising original and corrected sentence pairs along with several pre-processing stages, context-based error detection, correction ranking, and feedback generation. In addition, the findings indicate that local grammatical mistakes are easier to correct than complicated global structural mistakes, but correcting mistakes in order of importance still constitutes one of the greatest challenges that affect the efficiency of the system. Moreover, automatic feedback generation makes possible form-based and output-based learning through instantaneous learner self-awareness and self-correction. The findings suggest that NLP-based AGEC systems could be useful additional teaching tools in English language classes without the replacement of human teachers.

• PDF