EBMGC-GNF: Efficient Balanced Multi-View Graph Clustering via Good Neighbor Fusion

Area of Science:

• Data Science
• Machine Learning
• Artificial Intelligence

Background:

• Multi-view graph clustering is crucial for leveraging consistent structural information across datasets.
• Existing methods often struggle with effectively extracting and fusing credible neighbor information from multiple graph views.
• Achieving balanced cluster properties is essential for adapting to diverse data distributions.

Purpose of the Study:

• To propose an Efficient Balanced Multi-view Graph Clustering via Good Neighbor Fusion (EBMGC-GNF) model.
• To comprehensively extract credible consistent neighbor information using a Cross-view Good Neighbors Voting module.
• To introduce a novel balanced regularization term for adapting to different data distributions.

Main Methods:

• The EBMGC-GNF model utilizes a Cross-view Good Neighbors Voting module for neighbor information extraction.
• A novel balanced regularization term based on the p-power function is employed to adjust cluster balance.
• The optimization problem is efficiently solved using graph coarsening and an accelerated coordinate descent algorithm.

Main Results:

• Extensive experimental results demonstrate the effectiveness of the proposed EBMGC-GNF model.
• The model shows superior performance compared to state-of-the-art methods in most scenarios.
• Both the effectiveness and efficiency of the proposed approach are validated.

Conclusions:

• The EBMGC-GNF model provides an effective solution for multi-view graph clustering.
• The Cross-view Good Neighbors Voting module and balanced regularization term significantly improve clustering performance.
• The proposed method offers an efficient and effective approach for complex multi-view graph clustering tasks.