Clustering Methods for Identifying Key Framing Elements in Multimodal News Coverage of the Belt and Road Initiative: A Cross-platform Study

This study aims to determine the effectiveness of clustering strategies in the analysis of framing aspects in MMNs with regard to the Belt and Road Initiative (BRI). This paper seeks to understand how text, images and videos help in the construction of the BRI, a major geo-political and economic undertaking spear headed by China. To this end, this study intends to use a more sophisticated clustering approach that would allow identification of the main framing constituents and their interaction in the context of the specific media segments. The method adopted in the study combines text analysis, visual content analysis and extraction of frames from the videos as a way of capturing the dimensions of news coverage. The concept of multimodal framing is vital, especially when studying how such a large-scale global project like the BRI is framed. Thinking types involve textual factors like headlines or keywords and thematic stories together with visions like images and video clips addressed to find out the patterns and frames. The research applies NLP tools for textual content, image analysis for images, and video scribes for the classification of video clips. These methods combined provide a systematic way in which the presentation of content in multifaceted media can be studied, and give evidence on the continuity and disparity of the framing within platforms. It is only when one plays the different modes of communication against one another, that the study shows how each of them contributes to the construction of the public understanding of the BRI. In doing so, the study thus offers a holistic perspective of the nature and forms of the dominant frames alongside their differences across the media; a perspective that fills the gap left by existing research regarding how all the features of multimodal composition work together to shape the audience’s perception. This research contributes to the field of media and communication by proving the fact that how clustering technique is effective in multimodal analysis and how the current framework provides the foundation for future research related to the media framing that happens in the cross-media environment.