Multivariate Time Series Clustering with State Space Dynamical Modeling and Grassmann Manifold Learning: A Systematic Review on Human Motion Data

被引：0

作者：

Heo, Sebin ^{[1
]}

Teoh, Andrew Beng Jin ^{[2
]}

Yu, Sunjin ^{[3
]}

Oh, Beom-Seok ^{[1
]}

机构：

[1] Seoul Natl Univ Sci & Technol, Dept Appl Artificial Intelligence, Seoul 01811, South Korea

[2] Yonsei Univ, Sch Elect & Elect Engn, Seoul 03722, South Korea

[3] Changwon Natl Univ, Dept Culture Techno, Chang Won 51140, Gyeongsangnam D, South Korea

来源：

APPLIED SCIENCES-BASEL | 2025年 / 15卷 / 01期

关键词：

clustering; geodesic distance; Grassmann manifold; Karcher mean; kernel method; multivariate time series; tangent space; AFFINITY MATRIX; RECOGNITION; ALGORITHMS; SUBSPACES; NETWORK;

D O I：

10.3390/app15010043

中图分类号：

O6 [化学];

学科分类号：

0703 ;

摘要：

Multivariate time series (MTS) clustering has been an essential research topic in various domains over the past decades. However, inherent properties of MTS data-namely, temporal dynamics and inter-variable correlations-make MTS clustering challenging. These challenges can be addressed in Grassmann manifold learning combined with state-space dynamical modeling, which allows existing clustering techniques to be applicable using similarity measures defined on MTS data. In this paper, we present a systematic overview of Grassmann MTS clustering from a geometrical perspective, categorizing the methods into three approaches: (i) extrinsic, (ii) intrinsic, and (iii) semi-intrinsic. Consequently, we outline 11 methods for Grassmann clustering and demonstrate their effectiveness through a comparative experimental study using human motion gesture-derived MTS data.

引用

页数：19