Novel Clustering based on Discrete Morse Technique

Jian Ping Zhang, Xi Yu Liu, Yong Li

Abstract

A new clustering algorithm based on discrete Morse theory is proposed for cluster analysis in this paper. Firstly, an energy surface is defined on data set by Gaussian kernel functions. Secondly, a simplicial complex can be obtained by hull Triangulation on the energy surface. Finally, the optimization model based on discrete Morse theory is adopted to find cluster centers and clusters on a simplicial complex. It is a novel approach. The experimental results on some synthetic and UCI data sets have demonstrated that the new algorithm can discover clusters with arbitrary shapes and densities at different levels, moreover it can successfully divide data points overlapping into many meaningful clusters. The results show the feasibility and effectiveness of the new clustering algorithm.

Download


Paper Citation


in Harvard Style

Zhang J., Liu X. and Li Y. (2018). Novel Clustering based on Discrete Morse Technique.In Proceedings of the 7th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM, ISBN 978-989-758-276-9, pages 555-562. DOI: 10.5220/0006681805550562


in Bibtex Style

@conference{icpram18,
author={Jian Ping Zhang and Xi Yu Liu and Yong Li},
title={Novel Clustering based on Discrete Morse Technique},
booktitle={Proceedings of the 7th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,},
year={2018},
pages={555-562},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0006681805550562},
isbn={978-989-758-276-9},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 7th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,
TI - Novel Clustering based on Discrete Morse Technique
SN - 978-989-758-276-9
AU - Zhang J.
AU - Liu X.
AU - Li Y.
PY - 2018
SP - 555
EP - 562
DO - 10.5220/0006681805550562