传感器网络高维异常数据流动态挖掘算法设计  

Design of Dynamic Mining Algorithm for High Dimensional Abnormal Data Flow in Sensor Networks

作　　者：宁晓虹 NING Xiaohong(School of Information Technology and Engineering,Guangzhou College of Commerce,Guangzhou 510000,China)

机构地区：[1]广州商学院信息技术与工程学院,广东广州510000

出　　处：《传感器世界》2023年第11期34-39,共6页Sensor World

基　　金：基于网络自主学习平台的图形图像处理课程线上线下混合教学模式创新研究与实践(No.22GYB054)。

摘　　要：常规的传感网络异常数据流挖掘规则一般设置为独立形式,规则缺少交互过程,受到数据高维度的干扰,使得动态挖掘测算的覆盖范围受限制,且传感网络生成的数据量庞大,导致挖掘测算平均执行时间延长。因此,提出一种针对传感器网络高维异常数据流动态挖掘算法。根据当前的测定要求及标准,先进行异常数据流预处理,采用多层级的形式,打破动态挖掘测算覆盖范围受到的限制,制定多层级数据流挖掘规则。构建数据流链条交叉挖掘结构,同时建立深度集成异常数据流动态挖掘算法模型,采用动态增量测算实现数据挖掘算法。测试结果表明,所设计的算法测试组最终得出的挖掘测算平均执行时间被较好地控制在0.25 ms以下,准确率为95%,而其他传统方法的挖掘测算平均执行时间均在0.50 ms以上,准确率最高仅为87%。说明此次所设计的动态挖掘算法的效果更佳,精准度更高,具有实际的应用价值。The conventional rules for mining abnormal data streams in sensor networks are generally set in an independent form.The rules lack interactive processes and are affected by high-dimensional data interference,which limits the coverage of dynamic mining calculations.Additionally,the large amount of data generated by sensor networks results in an extension of the average execution time of mining calculations.Therefore,a dynamic mining algorithm for high-dimensional abnormal data streams in sensor networks is proposed.According to the current measurement requirements and standards,abnormal data flow preprocessing is carried out first,and a multi-level form is adopted to break the limitations of dynamic mining coverage and develop multi-level data flow mining rules.Construct a cross mining structure for the data flow chain,and establish a deep integration anomaly data flow dynamic mining algorithm model,using dynamic incremental calculation to achieve the data mining algorithm.The test results show that,the average execution time of mining calculation finally obtained by the designed algorithm test group is well controlled below 0.25 ms,the accuracy rate is 95%,while the average execution time of other traditional methods is above 0.50 ms,the accuracy rate is only 87%.It shows that the dynamic mining algorithm designed in this paper has better effect and higher accuracy,and has practical application value.

关 键 词：高维传感器 网络异常 异常数据 数据流挖掘 动态挖掘 算法设计 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]