A Comparative Study of Online Public Opinion Guiding Mechanisms Based on LDA Modeling and Sentiment Analysis

Under the background of deep integration of digitization and globalization, Internet public opinion governance has become a strategic high ground for shaping international discourse, online platforms have become a place for public opinion to spread and grow, the diversity of public opinion subjects and internationalization of the scope of dissemination have appeared, and public opinion guidance and governance play an important role in creating a clear cyberspace and shaping a harmonious social environment. In this study, we select regular political events and sudden political events with wide dissemination and great influence in the world as the field of analysis, and take the social media public opinion on the events at home and abroad as the object of study, and crawl posts and comments containing keywords through Python crawler and perform sentiment analysis and LDA theme model clustering to systematically deconstruct the paradigm differences between Chinese and Western public opinion governance. This study innovatively constructs a three-dimensional analysis framework of “cultural potential difference-governance effectiveness-communication path”, summarizes the cultural motivation and characteristic strategies of the public opinion guidance mechanism in Chinese and Western countries, and puts forward feasible suggestions for the improvement of China's public opinion guidance and governance system, which will also provide methodological support for the construction of the international communication discourse system with Chinese characteristics, and contribute to the cross-cultural adaptation of the global digital governance. It also contributes Chinese wisdom to the cross-cultural adaptation of global digital governance.