텍스트 마이닝을 활용한 한국어 학습자의 핵심 어휘 및 문법 사용 양상 연구 - 내러티브 텍스트를 중심으로 -

Abstract

This study aims to examine the usage patterns of linguistic expressions in narrative texts produced by foreign Korean learners compared with the text of Korean native speakers, and to reveal their characteristics objectively. To this end, the semantic network technique of text mining was used to differentiate it from the existing analysis methods. After extracting the words and grammatical elements considered statistically significant, we examined how the use of word, tense, and aspect differed according to the text structure. Korean native speakers and foreign Korean learners mostly developed their writings according to the structure of the introduction, events, and conclusion, and there was no significant difference. As a result of examining the relationship between key words and words for each structure, the two groups showed differences in content according to the topic in the introduction, events, and conclusion. Foreign learners displayed a characteristic of excessive use of first-person pronouns. In addition, as a result of examining the usage patterns of tenses and expressions in each structure, it was confirmed that foreign learners were unable to utilize the grammatical elements with tenses, aspect, and modalities as variedly as native Korean speakers did. This shows that the learner is accustomed to using grammar learned at the beginner level, and the acquisition related to tense and phase is still incomplete; hence, learning is required. This study is meaningful in that it attempted to computerize and explicitly identify linguistic features that were difficult to confirm through qualitative research methods to derive at results.