基于改进K-means算法的光通信数据异常检测预警方法

Optical communication data anomaly detection and early warning method based on improved k-means algorithm

在光通信系统的使用中,存在数据异常的情况,影响光通信质量,设计一种基于改进K-means算法的光通信数据异常检测预警方法。为解决容易陷入局部最优解与结果不稳定的问题,引入t-近邻距离思想改进K-means算法,实施光通信异常数据挖掘;通过自适应中值滤波方法实施光通信异常数据的去噪处理;基于动态时间规整与小波变换设计光通信数据异常检测模型;设计预警模式以实现光通信数据异常预警。测试设计方法的检测能力与性能,测试结果项目该方法能够明确检测到异常数据片段,异常检出率最高可达94.13%,灵敏度为0.99,前期增长比较显著。

In the use of optical communication systems,there are data abnormalities,which affect the quality of optical communication.A method of optical communication data abnormality detection and early warning based on improved K-means algorithm is designed.In order to solve the problem that it is easy to fall into the local optimal solution and the result is unstable,the idea of t-nearest neighbor distance is introduced,and the K-means algorithm is improved to implement the excavation of optical communication abnormal data.The method of adaptive median filter is used to de-noise the abnormal data of optical communication.Based on dynamic time warping and wavelet transform,an optical communication data anomaly detection model is designed and implemented.The early warning mode is designed to realize the abnormal early warning of optical communication data.Test the detection ability and performance of the design method.The test results show that the method can clearly detect abnormal data segments,and the abnormal detection rate can reach 94.13%,the sensitivity level was 0.99 and the early growth was relatively significant.