zbMATH — the first resource for mathematics

Similarity PCAs of multivariate time series based on an extended Frobenius norm. (Chinese. English summary) Zbl 1264.62055
Summary: The aim of this paper is to improve the efficiency of the similarity search of multivariate time series (MTSs). An extended Frobenius norm (EFroN) consisting of the eigenvalues and eigenvectors of the MTS covariance matrix is adopted for similarity Principal Component Analysis (PCA) of MTS data sets. Based on different similarity measurement methods, several similarity searching experiments are carried out to show the validity of the method. The results indicate that EFroN is superior in the recall precision as compared to the traditional similarity measures for MTS data sets.
62H25 Factor analysis and principal components; correspondence analysis
62M10 Time series, auto-correlation, regression, etc. in statistics (GARCH)