Abstract:A growing awareness of multi-view learning as an important component in data science and machine learning is a consequence of the increasing prevalence of multiple views in real-world applications, especially in the context of networks. In this paper we introduce a new scalability framework for multi-view subspace clustering. An efficient optimization strategy is proposed, leveraging kernel feature maps to reduce the computational burden while maintaining good clustering performance. The scalability of the algorithm means that it can be applied to large-scale datasets, including those with millions of data points, using a standard machine, in a few minutes. We conduct extensive experiments on real-world benchmark networks of various sizes in order to evaluate the performance of our algorithm against state-of-the-art multi-view subspace clustering methods and attributed-network multi-view approaches.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to solve the problems of computational efficiency and clustering performance of multi - view subspace clustering on large - scale datasets. Specifically: 1. **Challenges of multi - view subspace clustering**: - With the explosive growth of data from different sources, such as social media, sensor networks and online platforms, complex high - dimensional datasets have emerged. - These datasets usually contain multiple views of the same underlying data, and each view captures different aspects or perspectives. - Traditional clustering algorithms are designed for single - view data and face significant challenges when dealing with multi - view data, and it is difficult to effectively capture the complex relationships and structures across multiple views. 2. **Limitations of existing methods**: - Although some efficient methods have been proposed, such as anchor - based techniques to reduce the size of matrices used in optimization, tests show that existing methods often fail to achieve the required efficiency and clustering performance on large - scale multi - view datasets. - Computational complexity limits the wide application of multi - view subspace clustering in practical applications. 3. **Solutions proposed in this paper**: - A new scalable framework for multi - view subspace clustering is introduced. - An efficient optimization strategy is proposed, which uses kernel feature maps to reduce the computational burden while maintaining good clustering performance. - The algorithm can process large - scale datasets containing millions of data points within a few minutes on a standard machine. - Through extensive experimental verification, the algorithm performs excellently on real - world benchmark networks of different scales and outperforms existing multi - view subspace clustering methods and multi - view methods for attributed networks. 4. **Specific contributions**: - **Unified consensus subspace structure graph**: Integrate the data subspace structure graphs of multiple views into a unified consensus subspace structure graph, effectively capturing the complementary information between views. - **Efficient optimization strategy**: By taking advantage of the properties of kernel feature maps, the computational burden is significantly reduced while maintaining state - of - the - art clustering performance. - **Extensive experimental verification**: Extensive experiments including statistical significance tests are carried out to ensure the reproducibility and transparency of the results, enabling others to verify these results. In conclusion, this paper aims to solve the problems of computational efficiency and clustering performance of multi - view subspace clustering on large - scale datasets by introducing a new scalable framework and optimization strategy, thereby unlocking its full potential in practical applications.

Scalable Multi-view Clustering via Explicit Kernel Features Maps

Multi-View Clustering Via Simultaneously Learning Shared Subspace And Affinity Matrix

Efficient and Effective Deep Multi-view Subspace Clustering

Robust Kernelized Multiview Self-Representation for Subspace Clustering

Kernelized multi-view subspace clustering via auto-weighted graph learning

Unsupervised Multi-view Clustering by Squeezing Hybrid Knowledge from Cross View and Each View

Multi-View Data Fusion Oriented Clustering via Nuclear Norm Minimization

Scalable one-stage multi-view subspace clustering with dictionary learning

Deep Multiview Collaborative Clustering

An End-to-End Approach for Graph-based Multi-view Data Clustering

Multi-view clustering via matrix factorization assisted k -means

Fast and Scalable Semi-Supervised Learning for Multi-View Subspace Clustering

Exsavi: Excavating both sample-wise and view-wise relationships to boost multi-view subspace clustering

Efficient Multi-View -Means for Image Clustering

Multi-view low-rank sparse subspace clustering

Incomplete multi-view clustering via kernelized graph learning

Multi-view Clustering with Self-representation and Structural Constraint

Efficient Multi-View Clustering via Unified and Discrete Bipartite Graph Learning

Multi-view deep subspace clustering via level-by-level guided multi-level features learning

Multi-View Subspace Clustering Via A Global Low-Rank Affinity Matrix

Hierarchical Bipartite Graph Based Multi-View Subspace Clustering