基于秩2更新的多维数据流典型相关跟踪算法  

A Tracking Algorithm Based on Rank Two Modifications for Canonical Correlation Analysis of Multidimensional Data Streams

作　　者：杨静[1] 李文平[1] 张健沛[1] 

机构地区：[1]哈尔滨工程大学计算机科学与技术学院,黑龙江哈尔滨150001

出　　处：《电子学报》2012年第9期1765-1774,共10页Acta Electronica Sinica

基　　金：国家自然科学基金(No.61073043;No.61073041);黑龙江省自然科学基金(No.F200901);哈尔滨市科技创新人才研究专项资金(优秀学科带头人)(No.2011RFXXG015;No.2010RFXXG002);高等学校博士学科点基金(No.20112304110011)

摘　　要：现存的多维数据流典型相关分析(Canonical Correlation Analysis,简称CCA)算法主要是基于近似技术的求解方法,本质上并不是持续更新的精确算法.为了能在时变的环境中持续、快速而精确地跟踪数据流之间的相关性,本文提出一种多维数据流典型相关跟踪算法TCCA.该算法基于秩2更新理论,通过并行方式持续更新样本协方差矩阵的特征子空间,进而实现多维数据流典型相关的快速跟踪.理论分析及仿真实验结果表明,TCCA具有较好的稳定性、较高的计算效率和精度,可以作为基本工具应用于数据流相关性检测、特征融合、数据降维等数据流挖掘领域.Existing algorithms for canonical correlation analysis（CCA） of multidimensional data streams are mostly based on approximate techniques,but are not the precise algorithms for updates in essence.In this study,a novel canonical correlation analysis algorithm,called TCCA（Tracking CCA）,is proposed for tracking the correlations rapidly and accurately between two multidimensional data streams in the time-varying environments.By introducing the technique of rank two modifications to update the eigen-subspace of the sample covariance matrix in parallel,TCCA can rapidly track the correlations of data streams.Theoretical analysis and experimental results indicate that the TCCA algorithm has better stability,high computational efficiency and accuracy.It could be presented as a basic tool for correlation detection on data streams,feature fusion,dimension reduction and other areas of data streams mining.

关 键 词：多维数据流 典型相关分析 秩2更新 快速跟踪 特征子空间 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP311.1[自动化与计算机技术—计算机科学与技术]