【摘要】 大型数据传输中对其信息的异常预警,能够有效保障大规模数据驱动器安全平稳运行。对传输信息的异常快速预警,需要对异常预警指标构成不同的因素集进行计算,构造单因素动态模糊异常预警矩阵,完成数据传输信息的异常预警。传统方法结合情景分析,先检测出数据驱动器的信息传输自助性,但忽略了对异常预警矩阵的构建,导致预警精度偏低。提出基于自助式信息流模型的异常预警方法,采用十字转门模型来描述大型数据传输信息流,实时记录自助式信息流到sketch数据结构中的频率;引入动态模糊集,采用模糊综合评价方法对不同层次的异常预警指标构成不同的因素集进行计算;并赋予相应权重得到权重向量,构造单因素动态模糊异常预警矩阵,完成数据驱动器信息异常的快速预警。实验结果证明所提方法能够准确检测到大型数据传输信息追溯过程中的异常行为,判断异常预警等级。更多还原

【Abstract】 For the abnormal rapid alert of transmitted information,it is necessary to calculate different factor sets constituted by abnormal alert indicators.The traditional method ignores the construction of abnormal alert matrix,resulting in low accuracy.In this article,we focus on the method of abnormal early warning based on self-service information flow model.This method adopted the turnstile model to describe the large information flow of data transmission to record the frequency from self-service information flow to the sketch data structure.Then,we introduced dynamic fuzzy sets and used fuzzy comprehensive evaluation method to calculate different factor sets formed by different levels of abnormal early warning indicators.Meanwhile,we gave the corresponding weights to get weight vectors.Finally,we established the single-factor matrix of dynamic fuzzy abnormal early warning.Thus,we completed the rapid alert of abnormal information of data driver.Simulation results show that the proposed method can accurately detect the abnormal behavior in the process of information tracing of large data transmission information.It can judge the level of abnormal early warning.更多还原