登录    |    注册

您好,欢迎来到澳门新葡新京网赌!

澳门新葡新京首页> 澳门新葡新京网赌期刊 >本期导读>基于特征的时间序列信号表示方法

基于特征的时间序列信号表示方法

28    2020-05-27

¥0.50

全文售价

作者:沈培璐1, 汪朝海2, 钱源来1, 邵帅1, 周松飞1

作者单位:1. 中化兴中石油转运(舟山)有限公司, 浙江 舟山 316000;
2. 中国计量大学计量测试工程学院, 浙江 杭州 310018


关键词:时间序列数据;协方差矩阵;特征向量;动态时间规整;尺度不变特征变换


摘要:

为实现时间序列信号特征的预处理,提出一种基于特征协方差矩阵的时间序列信号表示方法。以一维时间序列数据为输入,计算每一采样时刻的点值、邻域差值、累加值、均差值、秩以及时间索引值等特征,组成点特征向量,将不同时刻的点特征向量依次带入黎曼流形空间,从而实现从一维采样数组到二维特征矩阵的转换;利用矩阵内含的时间序列局部和全局特征信息,计算二维特征矩阵的协方差矩阵,从而建立基于协方差矩阵的时间序列信号表示方法。为验证方法在特征表示方面的有效性,文中将其用于时间序列信号的相似度度量计算,实验结果表明,相比动态时间规整(DTW)、形状变换(ST)及其尺度不变特征变换(SIFT),文中算法总体分类性能为最佳,表明具有高效的特征表示性能。


Representation method of time series signals based on feature
SHEN Peilu1, WANG Chaohai2, QIAN Yuanlai1, SHAO Shuai1, ZHOU Songfei1
1. Sinochem Xingzhong Oil Staging (Zhoushan) Co., Ltd., Zhoushan 316000, China;
2. College of Metrology & Measurement Engineering, China Jiliang University, Hangzhou 310018, China
Abstract: In order to extract the features of vibration signals and other time series data, a time series representation method using feature covariance matrix is proposed. Taking one-dimensional time series data as input, calculating feature quantities such as point value, neighborhood difference value, accumulated value, average difference value, rank and time index value at each sampling time to form point feature vectors, and sequentially bringing the point feature vectors at different times into Riemannian manifold space, thus realizing conversion from one-dimensional sampling array to two-dimensional feature matrix; Using the local and global feature information of the time series contained in the matrix, the covariance matrix of the two-dimensional feature matrix is calculated to realize feature extraction. In the paper, the similarity measurement of time series signals is taken as an example for verification. Compared with dynamic time warping (DTW), shape transformation (ST) and scale invariant feature transformation (SIFT), indicating that it has more efficient feature representation performance.
Keywords: time series data;covariance matrix;feature vector;DTW;SIFT
2020, 46(5):13-18  收稿日期: 2019-01-23;收到修改稿日期: 2019-03-04
基金项目: 国家重点研发计划项目(2018YFF0214700);舟山市科技计划项目(2017c12036)
作者简介: 沈培璐(1986-),男,浙江舟山市人,助理工程师,硕士,主要从事储运系统信息化和设备安全监测
参考文献
[1] 孙友强. 时间序列数据挖掘中的维数约简与预测方法研究[D]. 合肥: 中国科学技术大学, 2014.
[2] 李海林, 郭崇慧. 基于形态特征的时间序列符号聚合近似方法[J]. 模式识别与人工智能, 2011, 24(5): 665-672
[3] 刘淑琴. 图像特征提取方法及其应用研究[D]. 西安: 西北大学, 2016.
[4] 田兴发. 面向视频篡改检测的特征提取算法研究[D]. 成都: 电子科技大学, 2016.
[5] ZHAO J, ITTI L. Classifying time series using local descriptors with hybrid sampling[J]. IEEE Trans Knowl Data Eng, 2016, 28(3): 623-637
[6] 王怀宝, 张杰, 张旭光, 等. 基于多特征协方差矩阵的图像匹配算法[J]. 计算机工程与设计, 2013, 34(9): 3190-3194
[7] TUZEL O, PORIKLI F, MEER P. Pedestrian detection via classification on riemannian manifolds[J]. IEEE Trans Pattern Anal Mach Intell, 2008, 30(10): 1713-1727
[8] YE L, KEOGH E. Time series shapelets: A new primitive for data mining[C]// Proceedings of 15th ACM SIGKDD international conference on knowledge discovery data mining, 2009: 947-956
[9] RATANAMAHATANA C A, KEOGH E. Three myths about dynamic time warping data mining [C]// Proceedings of the 2005 SIAM International Conference on Data Mining. 2005: 506-510.
[10] JASON L, ANTHONY B. Alternative quality measures for time series Shapelets[J]. Intelligent Data Engineering and Automated Learning, 2012, 74(35): 475-483
[11] YE L, KEOGH E. Time series Shapelets: A novel technique that allows accurate, interpretable and fast classification[J]. Data Mining and Knowledge Discovery, 2011, 22(1): 149-182
[12] LIN J, KEOGH E, LONARDI S, et al. A symbolic representation of time series, with implication for streaming algorithms[C]// Proceedings of Acm Sigmod Workshop on Research Issues in Data Mining & Knowledge Discovery. 2003: 2-11.
[13] 钟清流, 蔡自兴. 基于统计特征的时序数据符号化算法[J]. 计算机学报, 2008(10): 1857-1864
[14] 季海娟, 周从华, 刘志锋. 一种基于始末距离的时间序列符号聚合近似表示方法[J]. 计算机科学, 2018, 45(6): 216-221
[15] 王锐. 流形学习及其在图像集分类中的应用研究[D]. 无锡: 江南大学, 2018.
[16] ARSIGNY V, FILLARD P, PENNEC X, AYACHE N. Log-euclidean metrics for fast and simple calculus on diffusion tensors[J]. Magnetic Resonance in Medicine, 2006, 56(2): 411-421
[17] LOWE D G. Object recognition from local scale-invariant features [C]//Proceedings of the IEEE International Conference on Computer Vision. 1999, 2: 1150-1157.

Xml地图