.png){kind=icon}
.png){kind=icon}
Return
KDP-MHL: Key data point-aware multi-scale hypergraph learning framework for multivariate time series classification
.png){kind=icon}
PRE
AI
DOI：10.1016/j.knosys.2025.114620.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• This paper proposes a hypergraph learning framework KDP-MHL to improve MTSC performance by exploring complex relationships. • Establishing dynamic hypergraph structures can effectively extract high-order temporal associations. • The key data points selected in this paper, accounting for 25% of the original data, can still preserve the original temporal trends. • Comprehensive and rich class-specific information can improve the representation of complex temporal patterns. • The proposed method achieves SOTA performance, with an improvement of 3% in average accuracy.
Journal
K
IF:
7.6
Papers:
1.2W
Citations:
4.5W