Research on International Communication Power and the Construction of Foreign Discourse System

This study developed and validated a technology-enabled framework to optimize China's international communication. It aimed to address challenges of high cultural discount and weak technological adaptation hindering China's external discourse. The methodology involved an audience adaptation algorithm using K-means clustering to profile audiences and a bidirectional encoder representations from transformers (BERT) model for generating culturally sensitive messages. A random forest model was then built to predict communication effectiveness, followed by a controlled experiment on Twitter and TikTok data. Results showed the technology-optimized strategy more than doubled value recognition and interaction rates, with the prediction model achieving high accuracy (mean square error = 0.047, R^2 = 0.89). Sub-audience engagement rose by over 95%. This work provides a scalable technical solution that uses artificial intelligence to significantly enhance the precision and impact of international communication, enabling a shift from broadcasting to value-driven, adaptive engagement.