A Parallel Architecture for the Partitioning Around Medoids (PAM) Algorithm for Scalable Multi-Core Processor...

Area of Science:

• Computer Science
• Data Science
• Algorithm Optimization

Background:

• Clustering organizes unlabeled data into natural groups based on object similarity.
• The Partitioning Around Medoids (PAM) algorithm is a partitioning-based clustering method used in various fields.
• High time complexity limits PAM's application to large datasets and real-time systems.

Purpose of the Study:

• To propose a simple and scalable parallel architecture for the PAM algorithm.
• To reduce the running time and computational complexity of the PAM algorithm.
• To enable PAM's use in big data and real-time applications.

Main Methods:

• A parallel architecture that partitions data equally among multiple processing cores.
• Each core executes tasks simultaneously on its data subset and shares intermediate results.
• Implementation on multi-core processors or reconfigurable hardware (FPGA, MPSoCs).

Main Results:

• Exponential reduction in computational complexity with an increasing number of parallel cores.
• Linear speedup observed with more data points and uniform clusters.
• The proposed architecture achieves the same results as the original PAM algorithm with reduced complexity.

Conclusions:

• The developed parallel architecture effectively addresses the scalability and performance limitations of the PAM algorithm.
• The architecture is suitable for both big data processing and embedded real-time clustering applications.
• This work enhances the applicability of PAM in diverse data analysis scenarios.