Abstract:

The identification of innovation themes of key technologies in China's safety and emergency industry (SEI) and their evolutionary trends was investigated to improve the effectiveness of the digital and intelligent transformation of the industry. Key technology indicators (KIs) were proposed based on a large-scale dataset of 36 971 patents in the SEI filed from 2003 to 2024. Seven indicators spanning three dimensions, namely, technological, legal and strategic support, were selected to measure the key technologies. The XGBoost model was employed to identify key technological opportunities in the SEI. Topic modeling methods, including LDA, were further applied to extract innovation themes of key technologies and analyze their evolutionary trends. The results show that key digital and intelligent technologies in the SEI can be categorized into 10 thematic clusters, exhibiting logical correlation characteristics of "technology integration, proactive prevention, and coordinated response". The current technological development is reflected in three major trends of innovation themes: the evolution from data acquisition relying on conventional technologies to multimodal data fusion enabled by sensors, the transformation from focusing on emergency response to sudden events to emphasizing proactive monitoring and early warning, and the shift from focusing on traditional natural disaster prevention and control to resilient management of composite disasters.

Key words: digital and intelligent transformation, safety and emergency industry, key technology identification, evolutionary trend, latent dirichlet allocation (LDA)