A novel clustering algorithm inspired by membrane computing.

Area of Science:

• Computer Science
• Computational Biology
• Bio-inspired Computing

Background:

• P systems are a class of distributed parallel computing models.
• Clustering algorithms are essential for data analysis and pattern recognition.
• Existing clustering methods may face challenges with complex datasets and computational efficiency.

Purpose of the Study:

• To propose a novel clustering algorithm, the membrane clustering algorithm, inspired by tissue-like P systems.
• To leverage the parallel computing capabilities of P systems for effective data partitioning.
• To enhance object diversity and coevolution within the clustering process.

Main Methods:

• The membrane clustering algorithm utilizes a tissue-like P system with a loop structure of cells.
• Candidate cluster centers are represented as objects within cells and evolved using evolution rules.
• Communication rules establish a local neighborhood topology for coevolution and object diversity.

Main Results:

• The algorithm was evaluated on four artificial and six real-life datasets.
• Experimental results demonstrate the proposed algorithm's effectiveness.
• The membrane clustering algorithm shows superior or competitive performance against k-means and other evolutionary clustering algorithms.

Conclusions:

• The membrane clustering algorithm effectively utilizes the parallel computing advantages of tissue-like P systems.
• The local neighborhood topology enhances object coevolution and diversity, leading to improved clustering.
• The proposed algorithm offers a competitive and effective alternative for data clustering tasks.