텍스트 마이닝 기반 한국어 학습자 어휘 사용 양상 연구

Abstract

The purpose of this study is to measure ‘vocabulary diversity’, ‘vocabulary relationship complexity’, and ‘text coherence’ that appear in the written corpus of Korean learners by using the text mining technique to examine learners’ vocabulary usage patterns by proficiency level. ‘Vocabulary Diversity’, ‘Vocabulary Relationship Complexity’, and ‘Text Coherence’ were measured by targeting the corpus of 3-6 level learners corresponding to the middle and high level of Korean language education institutions using text mining techniques. As a result of the measurement, the higher of the learner’s mastery, the higher the value was measured. The vocabulary diversity increased to 0.0026 for the 3rd level, 0.0932 for the 4th level, 0.2382 for the 5th level, and 0.3658 for the 6th level, and the vocabulary relationship complexity increased to 0.2381 for the 3rd level, 0.3085 for the 4th level, 0.4325 for the 5th level, and 0.4899 for the 6th level. Text coherence rose to level 3 0.1717, level 4 0.1910, level 5 0.3210, and level 6 0.3684. As a result of conducting an analysis of variance to confirm whether these quantitative values were significant, the differences in the averages for each level of proficiency measured by applying the text mining technique were all significant and were consistent with qualitative analysis.