基于多目标决策的时间序列数据挖掘算法  

Time series data mining algorithm based on multi⁃objective decision

作　　者：贾晓强 JIA Xiaoqiang(School of Computer Science,Weinan Normal University,Weinan 714099,China)

机构地区：[1]渭南师范学院计算机学院,陕西渭南714099

出　　处：《电子设计工程》2021年第17期45-49,共5页Electronic Design Engineering

基　　金：陕西省教育厅科研基金资助项目(17JK0278);陕西教育厅科研计划项目(18JK0279);渭南师范学院科研基金资助项目(18YKS14);渭南师范学院教改项目(JG201926)。

摘　　要：为了解决当前算法挖掘时间序列数据过程中噪声对所得结果有较大影响的问题,该文采用多目标决策方法对时间序列数据进行挖掘。采用TOPSIS方法构造初始矩阵,并使用AHP方法确定权重,使用信息熵权重法对时间序列数据进行处理,采用K-means聚类方法对所得结果进行分类,并进行敏感性分析,发现TOPSIS的区间为[0.08,1],说明该模型鲁棒性强。因此用该模型进行不同类别的加速比测试,发现基于多目标决策理论的效果始终优于基于中心度和基于时空模型方法。In order to solve the problem that the current algorithm mining time series data will make the noise have a greater impact on the results,this paper makes an experiment of mining time series data using multi⁃objective decision⁃making method.TOPSIS method is used to construct the initial matrix and AHP method is used to determine the weight,and the information entropy method is used to process the time series data.K⁃means clustering method is used to classify the results,and sensitivity analysis is carried out.It is found that the interval of TOPSIS is[0.08,1],which shows that the model is robust.Therefore,using this model to test the acceleration ratio of different categories,it is found that the results based on multi⁃objective decision⁃making theory are always better than those based on centrality and spatiotemporal model.

关 键 词：多目标决策 时间序列 K-MEANS聚类 信息熵 加速比 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]