Research on Event Extraction and Event Relation Extraction for Strategic Operations Analysis

News contains all kinds of political, military, economic, social and other event information of interest to the strategic level, which can be extracted through intelligent methods, and can lay the foundation for the subsequent enhancement of strategic situational awareness capabilities. This paper proposes an information extraction framework and event ontology model for strategic operations research analysis based on the needs and characteristics of strategic operations research analysis. Using the method of “a small amount of manual annotation + fine-tuned large language model annotation”, we construct the information extraction dataset EfSOA for strategic operations research analysis, and propose the event extraction method based on the ABBSAC model and the event relation extraction method combining Roberta and Bi-FLASH-SRU. The experimental results show that the dataset constructed in this paper has complete elements, and the event extraction and event relation extraction methods are better than the comparative model, which enhances the ability of deep mining text information, expands the means of strategic operations research analysis, and effectively improves the comprehensiveness and scientificity of strategic decision-making.