Density propagation based adaptive multi-density clustering algorithm.

Area of Science:

• Data Science
• Machine Learning
• Artificial Intelligence

Background:

• Density-based clustering algorithms are sensitive to parameter selection, often requiring empirical tuning.
• Achieving optimal clustering performance typically relies on prior experimental knowledge.

Purpose of the Study:

• To propose a novel, adaptive density-based clustering algorithm that mitigates the need for manual parameter selection.
• To enhance the performance and applicability of density-based clustering for spatial data.

Main Methods:

• The Density Propagation based Adaptive Multi-density clustering (DPAM) algorithm was developed.
• DPAM employs a three-stage process: micro-cluster graph generation, density propagation with margin/cohesion redefinition, and regional density calculation.

Main Results:

• DPAM demonstrated superior performance compared to several state-of-the-art density clustering algorithms across various test cases.
• The algorithm's ability to function without parameter adjustment contributed to its promising results.

Conclusions:

• The proposed DPAM algorithm offers an effective, automated solution for density-based spatial data clustering.
• DPAM achieves high performance without manual parameter intervention, addressing a key limitation of existing methods.