From Barbarism to Civilization: The Evolving Image of China in International Cinema

This chapter applies unsupervised machine learning methods, including the LDA topic model, Word2Vec algorithm, and sentiment analysis, to examine the image of China in international cinema from the twentieth to the twenty-first centuries. Using 4,428 film plot synopses from the Internet Movie Database (IMDb), it explores thematic patterns, semantic associations, and sentiment variations, and employs Granger causality tests to link these with socioeconomic indicators such as Gross Domestic Product (GDP) and Foreign Direct Investment (FDI). The results show that China’s cinematic image has evolved from the “barbarian” to the “schemer” and finally to the “civilized great power,” reflecting global shifts in economic and political relations. This study demonstrates how large-scale text analysis can enrich cultural research within the social sciences by connecting media representations with broader historical and structural transformations.