面向高维异常数据挖掘的小波变换算法优化  

Optimization of Wavelet Transform Algorithm for High-Dimensional Anomaly Data Mining

作　　者：陈玲萍 杨呈永[2] CHEN Ling-ping;YANG Cheng-yong(Guilin Institute of Information Technology,Guilin Guangxi 541004,China;Guilin University of Technology,Guilin Guangxi 541004,China)

机构地区：[1]桂林信息科技学院,广西桂林541004 [2]桂林理工大学,广西桂林541004

出　　处：《计算机仿真》2025年第1期462-465,472,共5页Computer Simulation

基　　金：2022年度广西高等教育本科教学改革工程项目(2022JGA411);广西教育科学“十四五”规划2021年度专项课题(2021ZJY1489)。

摘　　要：为深度挖掘高维异常数据,获取准确数据信息,提出一种基于小波变换优化算法。利用小波变换算法充分挖掘数据的时空频率和变化特点,区分出低频和高频序列。判断数据异常值的真伪,确定挖掘极值对应的阈值区间,补全参数序列,滤除掉数据时频残差中噪声,测算傅里叶变换函数的时频序列,利用遗传算法优化小波变换,使算法能同时处理群体中的多个数据,获取异常数据挖掘的全局最优解,并确定数据位置与核心距离,完成高维异常数据的挖掘和判断。经实验证明所提优化后小波变换算法能有效挖掘并识别出高维异常数据,描述样本数据的变化波动,降低数据挖掘误差。In order to deeply mine high-dimensional abnormal data and obtain accurate data information,an optimization algorithm based on wavelet transform was proposed.Firstly,we used a wavelet transform algorithm to fully explore the spatiotemporal frequency and variation of data and distinguish between low-frequency and high-frequency sequences.Moreover,we judged the authenticity of data outliers,and then determined the threshold interval corresponding to the mining extreme value.Furthermore,we complemented the parameter sequence and filtered out the noise in the time-frequency residuals of the data.Meanwhile,we calculated the time-frequency sequence of the Fourier transform function and optimized the wavelet transform by genetic algorithm,so that the algorithm could simultaneously process multiple data in a group,thus obtaining the global optimal solution of abnormal data mining.Finally,we determined the data location and core distance,thus completing the mining and judgment for high-dimensional abnormal data.Experimental results show that the optimized algorithm can effectively mine and identify high-dimensional abnormal data,describe the changes and fluctuations of sample data,and reduce data mining errors.

关 键 词：高维异常数据 数据挖掘 小波变换 遗传算法 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]