Consistency-Induced Multiview Subspace Clustering.

Area of Science:

• Machine Learning
• Data Mining
• Computer Vision

Background:

• Multiview clustering is crucial for analyzing complex datasets.
• Existing methods struggle with high-dimensional data and view consistency.
• Exploiting shared information across views remains a challenge.

Purpose of the Study:

• To propose a novel multiview subspace clustering algorithm (CiMSC).
• To address limitations in handling high-dimensional data and view consistency.
• To improve the accuracy and robustness of multiview clustering.

Main Methods:

• CiMSC utilizes structural consistency (SC) and sample assignment consistency (SAC).
• SC ensures consistent connected components within each view's similarity matrix.
• SAC minimizes discrepancies in connected components across different views.

Main Results:

• CiMSC effectively handles high-dimensional data by learning subspace representations.
• The method achieves superior performance compared to 12 state-of-the-art algorithms.
• Demonstrated high accuracy, e.g., 98.06% on the BBCSport dataset.

Conclusions:

• CiMSC offers a robust framework for multiview subspace clustering.
• The proposed consistency-based approach enhances clustering performance.
• CiMSC provides an effective solution for complex, high-dimensional multiview data analysis.