kFuse: A novel density based agglomerative clustering

2025-05-09

AI Summary

Existing agglomerative clustering methods rely on multi-parameter tuning and suffer from unstable connectivity distance computation, leading to poor generalizability across datasets. This paper proposes kFuse, a density-driven agglomerative clustering method requiring only the desired number of final clusters as input. Its key contributions are: (1) automatic subcluster partitioning via natural neighbors, eliminating manual specification of local density thresholds; (2) a dual-criterion merging mechanism integrating boundary connectivity with density statistics (mean and variance), enhancing structural sensitivity and stability; and (3) an adaptive hierarchical merging rule that replaces conventional single-distance metrics. Extensive experiments on synthetic and real-world datasets demonstrate that kFuse significantly improves clustering accuracy and robustness—particularly under challenging conditions including complex cluster shapes, heterogeneous densities, and noise contamination.

Technology Category

• Machine Learning: Clustering
• Search and Optimization: Distributed Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Agglomerative clustering has emerged as a vital tool in data analysis due to its intuitive and flexible characteristics. However, existing agglomerative clustering methods often involve additional parameters for sub-cluster partitioning and inter-cluster similarity assessment. This necessitates different parameter settings across various datasets, which is undoubtedly challenging in the absence of prior knowledge. Moreover, existing agglomerative clustering techniques are constrained by the calculation method of connection distance, leading to unstable clustering results. To address these issues, this paper introduces a novel density-based agglomerative clustering method, termed kFuse. kFuse comprises four key components: (1) sub-cluster partitioning based on natural neighbors; (2) determination of boundary connectivity between sub-clusters through the computation of adjacent samples and shortest distances; (3) assessment of density similarity between sub-clusters via the calculation of mean density and variance; and (4) establishment of merging rules between sub-clusters based on boundary connectivity and density similarity. kFuse requires the specification of the number of clusters only at the final merging stage. Additionally, by comprehensively considering adjacent samples, distances, and densities among different sub-clusters, kFuse significantly enhances accuracy during the merging phase, thereby greatly improving its identification capability. Experimental results on both synthetic and real-world datasets validate the effectiveness of kFuse.

Problem

• Eliminates need for parameter tuning in agglomerative clustering
• Improves stability by redefining connection distance calculation
• Enhances accuracy via boundary connectivity and density similarity

Innovation

• Sub-cluster partitioning using natural neighbors
• Boundary connectivity via adjacent samples and shortest distances
• Density similarity assessment with mean density and variance