基于动态时间弯曲距离的灰关联度量方法及其应用  

Grey incidence measurement method based on dynamic time warping distance and its application

作　　者：张起荣[1] 刘歆[2] 

机构地区：[1]毕节学院信息工程学院,贵州毕节551700 [2]重庆邮电大学软件学院,重庆400065

出　　处：《重庆邮电大学学报（自然科学版）》2014年第5期700-705,共6页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：国家自然科学基金(61309014);重庆市基础与前沿研究计划项目(cstc2013jcyjA40009);毕节学院高层次人才科学研究项目(G2013006)~~

摘　　要：传统灰关联度量对于长度不同的序列采用删除较长序列数据、均值、GM(1,1)模型预测等方法进行补齐,导致不确定性信息增加,造成不必要的信息损失。将动态时间弯曲思想引入灰关联度量中,提出一种自适应序列长度的灰关联度。该方法无需补齐序列数据,利用序列间距离矩阵的最短路径作为相似判定依据。在此基础上,进一步构建了相应的灰关联度量方法,有效解决了序列数据长度不一致问题。为了证明方法的有效性,将其应用于分类算法设计中,取得了较为优异的测试结果。When the length of sequences is different, the traditional grey incidence measure mainly focuses on deleting the longer sequence data, mean or GM （ 1,1 ） model. It will increase the uncertainty information greatly. Based on dynamic time warping distance theory, a novel grey incidence measurement is proposed. It measures the degree of grey incidence by the shortest path in distance matrix of sequences and without padding other data. Moreover, the corresponding grey incidence measurement method based on dynamic time warping is constructed. It is an effective solution to solve the problem for different length between sequences. In order to prove its effectiveness, a corresponding classification algorithm is proposed. In the classification experiments, it achieves the better performance.

关 键 词：灰理论 动态时间弯曲 灰关联分析 灰关联度量 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]