Nuclear Pollution or Safe Discharge: Topic Evolution and Cognitive Analysis on Fukushima's Treated Radioactive Water

In the context of the Societies 5.0, a series of discussions on the emerging Fukushima treated radioactive water (FTRW) event was carried out, which has an impact on sustainable development in multiple fields including the economy, culture, and society. In order to comprehensively understand emerging topics and their evolution, explore the impact of people's cognition on their participation, and focus on people's attitudes and public participation in the FTRW event, we propose an evolution analysis framework (EAF) to analyze the massive multilingual comments and news collected from social media platforms in several countries. We design a multilingual topic extraction model (XLM-topic) to detect the patterns of topics and analyze their evolution. Potential relations between the FTRW event's elements are explored by relational reasoning based on a knowledge graph, which is established by entities and relations extracted from comments and news. Moreover, we predict the public attitudes and participation toward the FTRW event by utilizing our custom-designed public opinion cellular automata (POCA). The proposed POCA simulates the information dissemination, cognitive changes, and topic evolution among social groups in virtual spaces. It collaborates with XLM-topic to analyze trends in both physical and virtual spaces. Analysis results indicate that participants in different regions and countries have different attitudes and reactions toward the FTRW event, and the public's cognition on this event will interact with itself. Our study is conducive to promoting the integration and interaction of virtual space and physical space in the context of Societies 5.0, providing decision-making support for building a more harmonious and stable social environment.