.png){kind=icon}
.png){kind=icon}
Return
Robust Representation Learning for Clean Feature Discovery in Incomplete Multi-View Clustering
.png){kind=icon}
OA
AI
P
C
R
H
Y
DOI：10.1002/aisy.70462.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Graph-based incomplete multi-view clustering (IMVC) methods have garnered significant attention due to their ability to capture intrinsic data structures in the presence of missing views. However, existing graph learning techniques predominantly rely on pairwise relationships between samples, which limits their capacity to model the complex and nonlinear structures inherent in real-world incomplete data. To address these limitations, a novel approach, robust feature discovery in incomplete multi-view clustering (RIMVC), is introduced, which integrates learning-based techniques into graph representation models. Furthermore, to mitigate the effects of noisy data, the well-established robust principal component analysis (RPCA) framework is leveraged to recover clean feature representations. The learning power of neural networks is then incorporated to refine graph learning, thereby enhancing the representation of incomplete multi-view data. This approach yields high-quality graph representations that significantly improve downstream clustering tasks. Extensive experimental evaluations demonstrate that RIMVC outperforms state-of-the-art IMVC methods, showcasing its superior performance and robustness.
Keywords:
graph learning
incomplete multi-view clustering
multi-view clustering
similarity graph
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
6.1
Papers:
1.9K
Citations:
8.4K
Organization
Cited Papers