다항 로지스틱 회귀분석을 이용한 한국어 학습자의 언어권별 오류 분석

Analyzing the Error Types of Korean Language Learners Using Multinomial Logistic Regression

This study attempted to grasp the characteristics of errors by statistically analyzing the corpus of errors in large-scale learners. This study examined the characteristics of language-specific errors in the location, pattern, and level of errors in the learner’s corpus using the multinominal logistic regression model along with the occurrence rate. As a result, it was possible to identify characteristic error types with high or low probability of occurrence by student’s language. Through this study, the results that can be referred to in order to summarize the characteristic types of language-specific errors were presented. This study is meaningful in that it statistically searched for significant variables for each language region by comparing and analyzing the overall errors using large-scale learner balance corpus data.