Answers are grounded in this paper and cite their sources.
This study analyzes 5162 explanatory entries from the Hanja Etymology Dictionary using quantitative-linguistic methods. Each entry was coded for the presence of four descriptive elements: graphic-form, component-functional, semantic-development, and documentary-cultural. Because more than one element could appear in a single entry, the coding was conducted as multi-label annotation. A dominant label was also assigned to each entry to identify its main explanatory type. Term frequency–inverse document frequency (TF-IDF) analysis was used to examine the characteristic vocabulary of each descriptive element. Logistic regression and random forest were then applied to predict the dominant labels. Their overall accuracies were 0.845 and 0.874, respectively, although the performance differed across categories. The component-functional category showed the most stable classification, while the graphic-form and semantic-development categories were more frequently confused. Correspondence analysis showed the main relationships between the descriptive elements and their characteristic vocabulary, and the first two dimensions accounted for 86.7% of the total inertia. The results show that the explanatory entries cannot be divided into completely separate types. Different descriptive elements often occur together in a single entry. In particular, 54.5% of the entries contained both component-functional and semantic-development elements. This suggests that structural explanation and semantic development are frequently combined in the dictionary. The study therefore shows that multi-label annotation is useful for examining the structure of dictionary explanations and provides a basis for further corpus-based research on Hanja etymology dictionaries.
No takes yet. Share an insight, caveat, or question.
Im et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: