Research Article
Efficient Network Security Situation Assessment With Multi-Strategy DBO-SVR Hybrid Model
@INPROCEEDINGS{10.4108/eai.18-12-2025.2365269, author={Xuetao Du and Ling Chang and Xin Yan and Chen Zhang}, title={Efficient Network Security Situation Assessment With Multi-Strategy DBO-SVR Hybrid Model}, proceedings={Proceedings of the 13th International Conference on Identification, Information and Knowledge in the Internet of Things, IIKI 2025, 18-21 December 2025, Chengdu, China}, publisher={EAI}, proceedings_a={IIKI}, year={2026}, month={6}, keywords={Network security Situation assessment Multi-strategy DBO SVR machine}, doi={10.4108/eai.18-12-2025.2365269} }- Xuetao Du
Ling Chang
Xin Yan
Chen Zhang
Year: 2026
Efficient Network Security Situation Assessment With Multi-Strategy DBO-SVR Hybrid Model
IIKI
EAI
DOI: 10.4108/eai.18-12-2025.2365269
Abstract
Network security situation assessment (NSSA) plays a critical role in addressing dynamic and complex network threats, yet existing approaches face notable limitations: traditional model-driven methods lack adaptability to large-scale dynamic networks, data-driven Support Vector Regression (SVR) is highly sensitive to parameter tuning, and the original Dung Beetle Optimization (DBO) algorithm suffers from insufficient population diversity, unbalanced search capabilities, and proneness to local optima. For mitigating these drawbacks, we introduce a multi-strategy improved DBO (MIDBO) algorithm. This method fuses the chaotic elite opposition-based learning strategy, Levy flight strategy, and a modified spiral search mechanism, aiming to resolve the inherent limitations of the original DBO. Specifically, MIDBO is employed to optimize the kernel function parameters and penalty factors of SVR, constructing a hybrid MIDBO-SVR model for NSSA. Experimental results illustrate that the proposed model achieves more remarkable performance in assessment accuracy and convergence speed when compared to current methods such as SVR, APSO-SVR, and DBO-SVR.
