본문 바로가기
  • Home

A Study on an Early Diagnosis System for Children Language Developmental Disorders Using KoBERT

  • Journal of The Korea Society of Computer and Information
  • Abbr : JKSCI
  • 2025, 30(5), pp.51~58
  • Publisher : The Korean Society Of Computer And Information
  • Research Area : Engineering > Computer Science
  • Received : March 24, 2025
  • Accepted : April 29, 2025
  • Published : May 30, 2025

Won-Cheol Park 1

1국립공주대학교

Accredited

ABSTRACT

This study proposes an AI-based early diagnosis system for children's language developmental disorders using the Korean BERT (KoBERT) model. A dataset of 10,000 sentences was constructed, including typical expressions, autism spectrum disorder, stuttering, and communication disorders. The system was trained and evaluated to classify each type of disorder. Experimental results showed that the KoBERT-based system achieved the highest performance with an F1-score of 0.88, and the confusion matrix analysis confirmed accurate classification even between similar linguistic patterns. This study demonstrates that deep learning-based natural language processing (NLP) techniques can effectively support early diagnosis and intervention of language developmental disorders, contributing to clinical and educational applications.

Citation status

* References for papers published after 2023 are currently being built.