Emotional Semantic Decoding and Promotion Strategies of Brand Reputation from the Perspective of Multilingual Communication: An Empirical Study Based on Convergent Media Big Data

Brand reputation, as a core intangible asset for brands, faces challenges of "semantic distortion" and "emotional dislocation" in multilingual communication under the convergent media context. To address this issue, this study constructs a four-stage emotional semantic decoding framework of "data collection - word segmentation - semantic extraction - difference analysis" by integrating the BERT-BiLSTM-GRU model and cross-cultural communication theory. Empirical research is conducted using convergent media big data from three languages (Chinese, English, Thai) across platforms such as Weibo, YouTube, and LINE Today Thailand. The results reveal three key findings: first, there are significant dimensional differences in brand emotional expression across languages-Chinese focuses on social value (responsibility, credibility), English on economic value (innovation, competitiveness), and Thai on experience value (user experience); second, the core causes of cross-linguistic emotional semantic deviation are cultural context differences (42.3%), professional term translation errors (31.7%), and emotional expression habit differences (26.0%); third, a three-in-one strategy of "language adaptation - emotional resonance - scenario-specific" can effectively reduce emotional deviation and enhance brand reputation, which is verified by quasi-experiments. This study provides a technical framework and practical strategies for brand reputation management in multilingual communication scenarios.