中圖分類號：TP393文獻(xiàn)標(biāo)志碼：ADOI:10.19358/j.issn.2097-1788.2026.01.003 中文引用格式：高原， 汪辰瑞. 高噪聲日志攻擊源識別方法研究及實現(xiàn)［J］.網(wǎng)絡(luò)安全與數(shù)據(jù)治理，2026，45（1）：14-19. 英文引用格式：Gao Yuan， Wang Chenrui. Research on methods and systems for identifying highnoise log attack sources［J］.Cyber Security and Data Governance，2026，45（1）：14-19.
Research on methods and systems for identifying high-noise log attack sources
Gao Yuan1，2， Wang Chenrui1，3
1. Anhui Provincial Key Laboratory of Water Science and Smart Water Conservancy; 2. Anhui Dayu Water Conservancy Engineering Technology Co., Ltd.; 3. Anhui Provincial Construction Engineering Quality Supervision and Testing Station Co., Ltd.
Abstract： With the expansion of information system scale and the diversification of network attack methods, network security situation awareness platforms and other operation and support platforms generally suffer from problems such as alarm fatigue, high false alarm rates, and difficulty in attack attribution when facing massive heterogeneous logs. To address the challenges of attack source identification and threat attribution in highnoise log environments, this paper proposes a method for identifying attack sources in highnoise logs. This method uses a dynamic scoring model of attack source IPs based on multidimensional rules to achieve dynamic assessment and updating of the threat level of attack sources. Simultaneously, the system utilizes knowledge graphs to complete attack chain reconstruction and visualization analysis, improving the interpretability and handling efficiency of security incidents. Experimental results show that this method achieves a log compression rate of 99.6% on real log data in the water conservancy industry, reducing the false alarm rate to 8.3%, significantly improving security operation efficiency and response capabilities. The research results provide a feasible technical path for intelligent operation of industrylevel network security.
