Reweighted multi-view clustering with tissue-like P system.

Area of Science:

• Data Science
• Machine Learning
• Computational Biology

Background:

• Multi-view clustering is crucial for uncovering heterogeneous data information.
• Determining optimal weight distribution across different data views remains a significant challenge.
• Existing methods often struggle with computational efficiency and accurate view fusion.

Purpose of the Study:

• To propose a novel algorithm, Reweighted Multi-view Clustering with Tissue-like P system (RMVCP), for effective multi-view clustering.
• To address the challenge of view weight distribution in multi-view data.
• To enhance computational efficiency through parallel processing.

Main Methods:

• A two-step data processing approach involving self-representation for similarity matrix construction.
• View fusion techniques to obtain unified and updated similarity matrices.
• Integration of the algorithm within a tissue-like P system framework for improved computational parallelism.
• Application of Constrained Laplacian Rank (CLR) for direct clustering result generation.

Main Results:

• The RMVCP algorithm successfully fuses multiple data views into a unified similarity matrix.
• Direct clustering results are obtained without requiring additional post-processing steps.
• Experimental results demonstrate superior performance compared to state-of-the-art multi-view clustering algorithms.
• Significant improvements in computational efficiency were observed due to the P system integration.

Conclusions:

• RMVCP offers an effective solution for the view weight distribution problem in multi-view clustering.
• The algorithm achieves high clustering accuracy and enhanced computational efficiency.
• The proposed method provides a robust framework for analyzing heterogeneous data from multiple sources.