Comparison of Chinese Silk Culture and National Images in East Asian Media Reports Based on Multilingual Co-Occurrence Semantic Networks

Wenxiang Zhang, Sutian Xu, Fang Ouyang

Article
2025 / Volume 8 / Pages 881-903
Received 8 July 2025; Accepted 21 August 2025; Published 3 December 2025
https://doi.org/10.31881/TLR.2025.881

Abstract
To address the cognitive biases in the textile industry’s global narrative caused by cross-language semantic differences in the dissemination of Chinese silk culture and its modern innovations in East Asia, this paper proposes a dynamic cognitive graph modeling method based on the multilingual co-occurrence semantic network (M-CSN). This study constructs a textile-field corpus from media reports in China, Japan, South Korea, and North Korea from 2015-2023. The bidirectional encoder representations from transformers (BERT)-multilingual model is utilized to perform cross-language semantic alignment of silk textile terms. By combining the PageRank algorithm with sentiment weight embedding, the semantic association differences of core concepts such as “silk-innovation” and “silk-ecology” is quantitatively analyzed, focusing on the communication efficiency of modern textile technology. Research reveals that the semantic density index (SDI) of Japanese media has reached 8.2, with a focus on “dyeing technology”, whereas the emotional polarity value of Korean media with respect to smart textile innovation has reached 0.79, which is higher than the value of 0.65 for Chinese media. Furthermore, the co-occurrence weight of North Korea’s “trade data” reaches 0.73, and the annual growth rate of the co-occurrence frequency of South Korea’s “silk-virtual fitting” is 18.9%. These results demonstrate that the M-CSN can accurately locate cultural discount nodes, providing data support for formulating differentiated communication strategies for textile technologies and products. This framework effectively optimizes the global narrative path of silk culture, facilitating its transformation from “traditional symbols” to “modern values”.

Keywords
Chinese silk culture, East Asian media, textile technology propagation, multilingual co-occurrence semantic network, dynamic cognitive graph