摘要
提出了一种新型主从式警报融合机制,该机制依据一定的规则,在时间方向上对重复报警进行压缩;在空间方向上借助概率模糊认知图的表达、推理能力,对来自多个检测器的报警进行融合。理论分析和实验仿真均表明:该机制能有效地降低虚警和减少重复报警。
A new alerts fusion mechanism is given. In this model, repetitive alerts are compressed in temporal space based on some rules and alerts from several detectors are fused in spatial space based on PFCM. Both theoretic analysis and experiments prove that this model can decrease false positive and repetitive alerts.
出处
《计算机工程》
CAS
CSCD
北大核心
2005年第21期118-120,共3页
Computer Engineering
基金
国家"863"计划基金资助项目(2002AA4Z3420)
关键词
概率模糊认知图
警报融合
虚警
Probabilistic fuzzy cognitive maps
Alerts fusion
False positive
