Improving Density Peaks Clustering through GPU acceleration

Liu, Zhuojin; Gong, Shufeng; Su, Yuxuan; Wan, Changyi; Zhang, Yanfeng; Yu, Ge

doi:10.1016/j.future.2022.11.033

Cited by 6 publications

(1 citation statement)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Gong et al [31] parallelized DPC in a distributed setting and employed Voronoi diagrams to improve its efficiency. Liu et al [47] proposed a DPC algorithm for GPUs. Amagata and Hara [3] leveraged the kd-tree data structure to improve the efficiency of density computation and dependent point finding.…”

Section: Density Peaks Clustering (Dpc)mentioning

confidence: 99%

Faster Parallel Exact Density Peaks Clustering

Huang¹,

Yu²,

Shun³

2023

SIAM Conference on Applied and Computational Discrete Algorithms (ACDA23)

View full text Add to dashboard Cite

Clustering multidimensional points is a fundamental data mining task, with applications in many fields, such as astronomy, neuroscience, bioinformatics, and computer vision. The goal of clustering algorithms is to group similar objects together. Density-based clustering is a clustering approach that defines clusters as dense regions of points. It has the advantage of being able to detect clusters of arbitrary shapes, rendering it useful in many applications.In this paper, we propose fast parallel algorithms for Density Peaks Clustering (DPC), a popular version of density-based clustering. Existing exact DPC algorithms suffer from low parallelism both in theory and in practice, which limits their application to largescale data sets. Our most performant algorithm, which is based on priority search kd-trees, achieves O(log n log log n) span (parallel time complexity) for a data set of n points. Our algorithm is also work-efficient, achieving a work complexity matching the best existing sequential exact DPC algorithm. In addition, we present another DPC algorithm based on a Fenwick tree that makes fewer assumptions for its average-case complexity to hold.We provide optimized implementations of our algorithms and evaluate their performance via extensive experiments. On a 30core machine with two-way hyperthreading, we find that our best algorithm achieves a 10.8-13169x speedup over the previous best parallel exact DPC algorithm. Compared to the state-of-the-art parallel approximate DPC algorithm, our best algorithm achieves a 1.5-4206x speedup, while being exact.

show abstract

Section: Density Peaks Clustering (Dpc)mentioning

confidence: 99%