Mathematics, Год журнала: 2025, Номер 13(9), С. 1422 - 1422

Опубликована: Апрель 26, 2025

Multiview Clustering (MVC) plays a crucial role in the holistic analysis of complex data by leveraging complementary information from multiple perspectives, necessity era big data. Non-negative Matrix Factorization (NMF)-based methods have demonstrated their effectiveness and broad applicability clustering tasks, as they generate meaningful attribute distributions cluster assignments. However, existing shallow NMF approaches fail to capture hierarchical structures inherent real-world data, while deep ones overlook accumulation reconstruction errors across layers solely focusing on global loss function. To address these limitations, this study aims develop novel method that integrates an autoencoder-inspired structure into framework, incorporating layerwise error-correcting constraints. This approach can facilitate extraction features effectively mitigating error architectures. Additionally, repulsion-attraction manifold learning is incorporated at each layer preserve intrinsic geometric within The proposed model evaluated five multiview datasets, with experimental results demonstrating its capturing representations improving performance.

Язык: Английский