|
基于QSSVM的工业无线传感器网络异常数据检测分析
资源天下 2025/5/13 9:50:12
淮安生物工程高等职业学校,计算机工程系,江苏淮安 223200
摘要:工业的网络安全直接影响到经济效益,无线传感器网络(Wireless Sensor Network, WSN)异常数据处理显得尤为重要。因此设计了一种基于1/4超球面支持向量机(Quarter-Sphere support vector machines, QSSVM)方法的工业无线传感器网络异常数据检测方法,实现在线测试功能的异常检测算法。研究结果表明:增大视窗尺寸时,所需的运算时间也明显增加,算法精度也发生了连续提升。检测能力则会随采样维数的增加而发生大幅提升。相对K-means算法,QSSVM检测率明显增加,可以达到99.5%以上。该研究能够满足工业大量数据处理需求,具有很高的应用价值。
关键词:工业安全;无线传感器网络;异常数据;支持向量机
中图分类号:TP393 文献标识码:A
Abnormal data detection and analysis of industrial wireless sensor networks based on QSSVM
Tan Hongmei
Department of Computer Engineering, Huaian Higher Vocational School of Biological Engineering, Huaian 223200, China
Abstract: The industrial Network security directly affects the economic benefits, Wireless Sensor Network (WSN) abnormal data processing is particularly important. Therefore, an anomaly detection method based on Quarter-Sphere support vector machines (QSSVM) is designed to realize the anomaly detection algorithm of online testing function in industrial wireless sensor networks. The results show that when the window size is increased, the operation time is obviously increased, and the algorithm accuracy is continuously improved. The detection capability will be greatly improved with the increase of sampling dimension. Compared with K-means algorithm, the detection rate of QSSVM is significantly increased, which can reach more than 99.5%. This research can meet the needs of large amount of data processing in industry and has high application value.
Key words: industrial safety; Wireless sensor network; Abnormal data; Support vector machine
0 引言
目前,随着信息通信技术的快速发展,物联网技术也获得了越来越广泛的应用,对人们的日常生产与社会发展都起到了重要支持[1]。然而,考虑到在实际使用环境中存在多种复杂因素,且不同传感器之间进行数据传输的方式与系统结构也有很大的差别,因此,在发生故障的时候,数据会立刻呈现出异常特征,从而对整个系统造成很大冲击,导致很多精确参考数据难以获取,从而会对最后策略制定过程产生影响[2-3]。
当出现个别异常情况时,可以根据集群的不同进行分类[4]。还有学者提出了新的检测技术来实现对复杂环境条件下的高维数据集实施检测,这也成为了当前的一种非常有效检测技术[5]。相关方面的研究吸引了很多的学者,取得了一定的研究成果。车丽娜等[6]提出了基于滑动窗口和置信度的WSN异常检测算法,引入多通信半径划分最佳邻域,根据节(未完,下一页)
附件下载:基于QSSVM的工业无线传感器网络异常数据检测分析
|