A parameter-free graph reduction for spectral clustering and SpectralNet

被引:4
|
作者
Alshammari, Mashaan [1 ]
Stavrakakis, John [2 ]
Takatsuka, Masahiro [2 ]
机构
[1] Univ Hail, Coll Comp Sci & Engn, Hail 81411, Saudi Arabia
[2] Univ Sydney, Sch Comp Sci, Sydney, NSW 2006, Australia
关键词
Spectral clustering; SpectralNet; Graph reduction; Local scale similarity; SPARSIFICATION;
D O I
10.1016/j.array.2022.100192
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Graph-based clustering methods like spectral clustering and SpectralNet are very efficient in detecting clusters of non-convex shapes. Unlike the popular k-means, graph-based clustering methods do not assume that each cluster has a single mean. However, these methods need a graph where vertices in the same cluster are connected by edges of large weights. To achieve this goal, many studies have proposed graph reduction methods with parameters. Unfortunately, these parameters have to be tuned for every dataset. We introduce a graph reduction method that does not require any parameters. First, the distances from every point p to its neighbors are filtered using an adaptive threshold to only keep neighbors with similar surrounding density. Second, the similarities with close neighbors are computed and only high similarities are kept. The edges that survive these two filtering steps form the constructed graph that was passed to spectral clustering and SpectralNet. The experiments showed that our method provides a stable alternative, where other methods' performance fluctuated according to the setting of their parameters.
引用
收藏
页数:9
相关论文
共 50 条
  • [1] A parameter-free similarity graph for spectral clustering
    Inkaya, Tulin
    EXPERT SYSTEMS WITH APPLICATIONS, 2015, 42 (24) : 9489 - 9498
  • [2] A novel method of spectral clustering in attributed networks by constructing parameter-free affinity matrix
    Berahmand, Kamal
    Mohammadi, Mehrnoush
    Faroughi, Azadeh
    Mohammadiani, Rojiar Pir
    CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS, 2022, 25 (02): : 869 - 888
  • [3] A novel method of spectral clustering in attributed networks by constructing parameter-free affinity matrix
    Kamal Berahmand
    Mehrnoush Mohammadi
    Azadeh Faroughi
    Rojiar Pir Mohammadiani
    Cluster Computing, 2022, 25 : 869 - 888
  • [4] Spectral methods for graph clustering - A survey
    Nascimento, Maria C. V.
    de Carvalho, Andre C. P. L. F.
    EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2011, 211 (02) : 221 - 231
  • [5] Connected graph decomposition for spectral clustering
    Tong, Tao
    Zhu, Xiaofeng
    Du, Tingting
    MULTIMEDIA TOOLS AND APPLICATIONS, 2019, 78 (23) : 33247 - 33259
  • [6] Connected graph decomposition for spectral clustering
    Tao Tong
    Xiaofeng Zhu
    Tingting Du
    Multimedia Tools and Applications, 2019, 78 : 33247 - 33259
  • [7] Reeb graph computation through spectral clustering
    Ma, Teng
    Wu, Zhuangzhi
    Luo, Pei
    Feng, Lu
    OPTICAL ENGINEERING, 2012, 51 (01)
  • [8] REVISITING FAST SPECTRAL CLUSTERING WITH ANCHOR GRAPH
    Wang, Cheng-Long
    Nie, Feiping
    Wang, Rong
    Li, Xuelong
    2020 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, 2020, : 3902 - 3906
  • [9] Survey of spectral clustering based on graph theory
    Ding, Ling
    Li, Chao
    Jin, Di
    Ding, Shifei
    PATTERN RECOGNITION, 2024, 151
  • [10] Towards for Using Spectral Clustering in Graph Mining
    Ait El Mouden, Z.
    Moulay Taj, R.
    Jakimi, A.
    Hajar, M.
    BIG DATA, CLOUD AND APPLICATIONS, BDCA 2018, 2018, 872 : 144 - 159