Multi-Source Evidence Conflict Fusion for Network Security Situation Assessment

Network threats are developing rapidly, and conventional situation assessment approaches encounter problems like multi-source heterogeneous evidence conflict and intensive real-time requirements. In this paper, based on multi-source evidence conflict fusion, we put forward a network security situation assessment approach, which can represent heterogeneous evidence with tensor decomposition, design attribute-weighted conflict coefficients to measure evidence inconsistency, and introduce dynamic weight modification approaches to evidence fusion. Experimental outcomes prove that the new approach can handle evidence conflict well and significantly enhance the accuracy and reliability of situation assessment.