Identify Significant Phenomenon-Specific Variables for Multivariate Time Series

Identify Significant Phenomenon-Specific Variables for Multivariate Time Series
复制标题

DOI:
10.1109/tkde.2019.2934464
复制
发表时间:
2019-08
影响因子:
8.9
通讯作者:
Yifan Hao;H. Cao;A. Mueen;S. Brahma
Yifan Hao;H. Cao;A. Mueen;S. Brahma
中科院分区:
计算机科学2区
文献类型:
--
作者:
Yifan Hao;H. Cao;A. Mueen;S. Brahma

文献摘要

相似文献

多变量时间序列(MTS)是在研究科学现象或监测系统健康时为不同变量收集的,其中每个时间序列记录一个变量在一段时间内的值。在不同的变量中,通常只有少数变量对特定现象有显着贡献。此外,对不同现象有重大影响的变量往往不同。我们将导致不同现象发生的不同变量表示为现象特定变量(PV)。在本文中,我们制定了一个新的问题,从MTS数据集识别重要的PV。为了分析MTS数据,已经广泛地研究了特征提取技术。然而,他们中的大多数识别一个数据集的重要的全球功能,并没有利用时间序列的时间顺序。为了解决这个新引入的问题,我们提出了一个解决方案框架CNN$_{mts}$mts-X,它是卷积神经网络(CNN)的一个新变体,可以嵌入其他特征提取技术(如X)。此外,我们设计了一个CNN$_{mts}$mts-LR方法,该方法在CNN$_{mts}$mts-X框架中实现了一种新的特征识别方法(LR)作为X。LR方法利用线性判别分析(LDA)和随机森林(RF)。我们在五个真实的数据集上的广泛实验表明,CNN$_{mts}$mts-LR方法比其他几种基线方法表现出更好的性能。使用从CNN$_{mts}$mts-LR中发现的30%的PV,分类可以实现比使用所有变量更好或类似的性能。
Multivariate time series (MTS) are collected for different variables in studying scientific phenomena or monitoring system health where each time series records the values of one variable for a time period. Among the different variables, it is common that only a few variables contribute significantly to a specific phenomenon. Furthermore, the variables contributing significantly to different phenomena are often different. We denote the different variables that contribute to the occurrences of different phenomena as Phenomenon-specific Variables (PVs). In this paper, we formulate a novel problem of identifying significant PVs from MTS datasets. To analyze MTS data, feature extraction techniques have been extensively studied. However, most of them identify important global features for one dataset and do not utilize the temporal order of time series. To solve the newly introduced problem, we propose a solution framework, CNN$_{mts}$mts-X, which is a new variant of the Convolutional Neural Networks (CNN) and can embed other feature extraction techniques (as X). Furthermore, we design a CNN$_{mts}$mts-LR method that implements a new feature identification approach (LR) as X in the CNN$_{mts}$mts-X framework. The LR method leverages both Linear Discriminant Analysis (LDA) and Random Forest (RF). Our extensive experiments on five real datasets show that the CNN$_{mts}$mts-LR method has exhibited much better performance than several other baseline methods. Using 30 percent of the PVs discovered from the CNN$_{mts}$mts-LR, classifications can achieve better or similar performance than using all the variables.