대한언어학회 전자저널

Lu, Xinghua & Feng Xiaoying. (2025). Comparative study of Korean news articles in South Korea, China and North Korea based on regression analysis: Taking Korean characters as a case. The Linguistic Association of Korea Journal, 33(2), 27-49. This study focuses on the Korean characters appearing in Korean-language news texts from South Korea, China, and North Korea. Using regression analysis, a functional model of character frequency distribution curves was established, and the distribution patterns were comparatively analyzed. The results show that both unsegmented linear and nonlinear regression models performed poorly, whereas segmented nonlinear regression models significantly improved the fit. Among the three countries, North Korea exhibited the lowest character diversity, with a small number of high-frequency characters dominating the text, followed by South Korea and China. Although the overall shapes of the character frequency distribution curves were similar across the three countries, differences were observed in the rankings of specific characters, and each country featured some unique rare characters. These findings are expected to provide valuable reference for comparative studies on Korean-language statistics among the three countries.

# 한국(조선)어 문자(Korean character) # 신문기사(newspaper article) # 회귀분석(regression analysis) # 모델(function model) # Zipf 법칙(Zipf’s Law)

• 곽충구. (2001). 남북한 언어 이질화와 그에 관련된 몇 문제. 새국어생활, 11(1), 5-27.
• 심지영. (2014). 중국조선어 ‘먹어 못 내다’ 식 부정문에 대한 일고찰. 언어와 문화, 10(1), 127-151.
• 崔荣一, 赵雪. (2017). 齐普夫定律对朝鲜语适用性的测定. 中文信息学报, 31(5), 81-84, 91.
• 韩普, 路高飞, 王东波. (2012). 基于最大似然估计方法的齐普夫定律检验. 情报理论与实 践, 35(11), 6-11.
• 何晓群, 刘文卿. (2019). 应用回归分析. 北京: 中国人民大学出版社.
• 朴美玉. (2014). 延边朝鲜语词汇变异研究——以朝鲜语汉字词和外来词的使用为例. 云南 师范大学学报(哲学社会科学版), 46(4), 11-21.
• 朱曦, 李旸. (2014). 基于齐普夫定律的我国城市规模分布实证研究. 现代经济信息, 07, 3-4.