As global e-commerce expands, cross-cultural consumer sentiment analysis using native-language deep learning remains underexplored in online retail contexts. This study analyzes 33,464 online wine reviews from China (JD.com) and Italy (Amazon.it) to examine how culturally embedded review patterns are associated with consumer emotional expression in digital environments. By integrating Self-Determination Theory (SDT) and Hofstede's cultural dimensions within a Transformer-based natural language processing (NLP) framework, the study links theoretical constructs to observable sentiment patterns. The results reveal a systematic divergence between platform-generated star ratings and underlying textual sentiment. Chinese consumers emphasize functional and risk-related attributes, whereas Italian consumers focus more on experiential, social and aesthetic dimensions, suggesting that online reviews may serve as indicators of culturally embedded evaluation tendencies rather than purely platform-driven behaviors. Building on these findings, this study develops a cross-cultural digital sentiment matrix that conceptualizes consumer behavior along two dimensions: motivation and expression. This framework provides a transferable analytical reference for interpreting cross-cultural consumer sentiment and supporting data-driven retail decision-making.
Liu et al. (Tue,) studied this question.