面向不确定性多数据流异常检测的数学模型
A Mathematical Model for Anomaly Detection in Uncertain Multiple Data Streams
张学叶 1林永强2
作者信息
- 1. 重庆移通学院数理教学部,重庆 401520
- 2. 西南政法大学经济学院,重庆 401120
- 折叠
摘要
随着互联网技术的快速发展,数据流的应用日益普遍,通信平台对多数据流进行异常检测的需求也逐步增长.为了解决当前异常检测算法准确率低、特征提取难等问题,提出了一种基于网格化的多数据流异常检测算法.算法首先提取不确定性多数据流的特征,通过分析数据流分布状态,从而提取异常数据;然后采用网格化的方法对多数据流进行划分,通过计算网格异常因子从而提取异常数据,达到异常检测的效果;最后针对异常数据,通过对变量因素进行关联性分析,降低误检率,提升异常检测的准确率.实验结果表明,所提算法在异常检测精确度方面提升了约 4%,漏检率降低了至少 3%,误检率降低了 8%以上,有效的提高了异常检测的精确度,降低了异常数据流对工作及生活带来的负面影响.
Abstract
With the rapid development of Internet technology,the application of data streams is becoming more and more popular,and the demand of communication platforms for anomaly detection of multiple data streams is also growing.In order to solve the problems of low accuracy and difficult feature extraction of current anomaly detection al-gorithms,this paper proposes a grid based multi data stream anomaly detection algorithm.The algorithm first extracts the features of uncertain multi data streams,and by analyzing the distribution status of the data streams,abnormal data is extracted;Then,a grid based method is used to partition multiple data streams,and abnormal data is extracted by calculating the grid anomaly factor,achieving the effect of anomaly detection;Finally,for abnormal data,correlation a-nalysis of variable factors is conducted to reduce false positives and improve the accuracy of anomaly detection.The experimental results show that the algorithm proposed in this paper improves the accuracy of anomaly detection by a-bout 4%,reduces the rate of missed detection by at least 3%,and reduces the false detection rate by more than 8%.It effectively improves the accuracy of anomaly detection and reduces the negative impact of abnormal data flow on work and life.
关键词
多数据流/异常检测/数学模型/异常因子Key words
Multiple data streams/Abnormal detection/Mathematical model/Abnormal factor引用本文复制引用
基金项目
重庆市教委科学技术研究计划(KJ1600403)
出版年
2024
NETL
NSTL
万方数据