@article{ART003364230},
author={Byeong-chan Lee},
title={A High-Density Compressed Text Topic Modeling Scheme Based on Large Language Models: Focusing on Classical Chinese Poetry},
journal={Journal of The Korea Society of Computer and Information},
issn={1598-849X},
year={2026},
volume={31},
number={7},
pages={141-150}
TY - JOUR
AU - Byeong-chan Lee
TI - A High-Density Compressed Text Topic Modeling Scheme Based on Large Language Models: Focusing on Classical Chinese Poetry
JO - Journal of The Korea Society of Computer and Information
PY - 2026
VL - 31
IS - 7
PB - The Korean Society Of Computer And Information
SP - 141
EP - 150
SN - 1598-849X
AB - To resolve over-segmentation in modeling highly compressed texts like Classical Chinese poetry, we propose High-Density Compressed Text Topic Modeling (HDCTM). Unlike traditional models that extract keywords post-clustering, HDCTM utilizes a Large Language Model (LLM) at the pre-embedding stage. It extracts structured 3-axis metadata (Subject, Type, Emotion) as semantic anchors, which are fused with raw texts to construct a knowledge-augmented embedding space. The architecture executes a hybrid two-track process: top-down structural integration via entry-point sample extraction, and bottom-up contextual discovery driven by micro-level multi-sampling and macro-level geometric centroid verification. Evaluated on 5,247 classical poems, HDCTM successfully resolves semantic fragmentation, establishing a robust, explainable analytical framework for computational humanities.
KW - Topic Modeling;Large Language Models;Classical Chinese Poetry;Compressed Text;Multi-sampling;Centroid Verification;HDBSCAN;Computational Humanities
DO -
UR -
ER -
Byeong-chan Lee. (2026). A High-Density Compressed Text Topic Modeling Scheme Based on Large Language Models: Focusing on Classical Chinese Poetry. Journal of The Korea Society of Computer and Information, 31(7), 141-150.
Byeong-chan Lee. 2026, "A High-Density Compressed Text Topic Modeling Scheme Based on Large Language Models: Focusing on Classical Chinese Poetry", Journal of The Korea Society of Computer and Information, vol.31, no.7 pp.141-150.
Byeong-chan Lee "A High-Density Compressed Text Topic Modeling Scheme Based on Large Language Models: Focusing on Classical Chinese Poetry" Journal of The Korea Society of Computer and Information 31.7 pp.141-150 (2026) : 141.
Byeong-chan Lee. A High-Density Compressed Text Topic Modeling Scheme Based on Large Language Models: Focusing on Classical Chinese Poetry. 2026; 31(7), 141-150.
Byeong-chan Lee. "A High-Density Compressed Text Topic Modeling Scheme Based on Large Language Models: Focusing on Classical Chinese Poetry" Journal of The Korea Society of Computer and Information 31, no.7 (2026) : 141-150.
Byeong-chan Lee. A High-Density Compressed Text Topic Modeling Scheme Based on Large Language Models: Focusing on Classical Chinese Poetry. Journal of The Korea Society of Computer and Information, 31(7), 141-150.
Byeong-chan Lee. A High-Density Compressed Text Topic Modeling Scheme Based on Large Language Models: Focusing on Classical Chinese Poetry. Journal of The Korea Society of Computer and Information. 2026; 31(7) 141-150.
Byeong-chan Lee. A High-Density Compressed Text Topic Modeling Scheme Based on Large Language Models: Focusing on Classical Chinese Poetry. 2026; 31(7), 141-150.
Byeong-chan Lee. "A High-Density Compressed Text Topic Modeling Scheme Based on Large Language Models: Focusing on Classical Chinese Poetry" Journal of The Korea Society of Computer and Information 31, no.7 (2026) : 141-150.
