Analysis of personal extreme emotional perception based on open source intelligence

With the development of the network, extreme events emerge one after another, posing a serious threat to urban public safety, while the existing early warning methods have problems such as lag and insufficient coverage. This paper uses open source intelligence to construct a multi-dimensional personal extreme emotion perception analysis framework, integrates microblog text, interactive communication and structured risk indicators from multiple sources, combines sentiment analysis, graph neural network and machine learning risk assessment, and proposes an early warning system for personal extreme emotional risk based on IBR model, which provides low-cost, timely, and practical intelligence support and decision-making assistance for urban safety management.