Adaptive Graph K-Means

被引:0
作者
Pei, Shenfei [1 ]
Sun, Yuanchen [2 ]
Nie, Feiping [3 ]
Jiang, Xudong [4 ]
Zheng, Zengwei [1 ]
机构
[1] Hangzhou City Univ, Sch Comp & Comp Sci, Hangzhou 310015, Zhejiang, Peoples R China
[2] Zhejiang Univ, Coll Comp Sci & Technol, Hangzhou 310027, Zhejiang, Peoples R China
[3] Northwestern Polytech Univ, Sch Comp Sci, Xian 710072, Shaanxi, Peoples R China
[4] Nanyang Technol Univ, Sch Elect & Elect Engn, Singapore 639798, Singapore
基金
中国国家自然科学基金;
关键词
Machine learning; Clustering; Graph-based; k-means; Computational efficiency;
D O I
10.1016/j.patcog.2024.111226
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Clustering large-scale datasets has received increasing attention recently. However, existing algorithms are still not efficient in scenarios with extremely large number of clusters. To this end, Adaptive Graph K-Means (AGKM) is proposed in this work. Its idea originates from k-means, but it operates on an adaptive k-Nearest Neighbor (k-NN) graph instead of data features. First, AGKM is highly efficient for processing datasets where both the numbers of samples and clusters are very large. Specifically, the time and space complexity are both linear w.r.t the number of samples and, more importantly, independent to the cluster number. Second, AGKM is designed for balanced clusters. This constraint is realized by adding a regularization term in loss function, and a simple modification of the graph in optimization algorithm, which does not increase the computational burden. Last, the indicator and dissimilarity matrices are learned simultaneously, so that the proposed AGKM obtains the final partition directly with higher efficacy and efficiency. Experiments on several datasets validate the advantages of AGKM. In particular, over 29X and 46X speed-ups with respect to k-means are observed on the two large-scale datasets WebFace and CelebA, respectively.
引用
收藏
页数:11
相关论文
共 50 条
  • [21] Evaluation of modified adaptive k-means segmentation algorithm
    Taye Girma Debelee
    Friedhelm Schwenker
    Samuel Rahimeto
    Dereje Yohannes
    Computational Visual Media, 2019, 5 : 347 - 361
  • [22] Adaptive Explicit Kernel Minkowski Weighted K-means
    Aradnia, Amir
    Haeri, Maryam Amir
    Ebadzadeh, Mohammad Mehdi
    INFORMATION SCIENCES, 2022, 584 : 503 - 518
  • [23] Sparse probabilistic K-means
    Jung, Yoon Mo
    Whang, Joyce Jiyoung
    Yun, Sangwoon
    APPLIED MATHEMATICS AND COMPUTATION, 2020, 382
  • [24] Robust trimmed k-means
    Dorabiala, Olga
    Kutz, J. Nathan
    Aravkin, Aleksandr Y.
    PATTERN RECOGNITION LETTERS, 2022, 161 : 9 - 16
  • [25] Transformed K-means Clustering
    Goel, Anurag
    Majumdar, Angshul
    29TH EUROPEAN SIGNAL PROCESSING CONFERENCE (EUSIPCO 2021), 2021, : 1526 - 1530
  • [26] K*-Means: An Effective and Efficient K-means Clustering Algorithm
    Qi, Jianpeng
    Yu, Yanwei
    Wang, Lihong
    Liu, Jinglei
    PROCEEDINGS OF 2016 IEEE INTERNATIONAL CONFERENCES ON BIG DATA AND CLOUD COMPUTING (BDCLOUD 2016) SOCIAL COMPUTING AND NETWORKING (SOCIALCOM 2016) SUSTAINABLE COMPUTING AND COMMUNICATIONS (SUSTAINCOM 2016) (BDCLOUD-SOCIALCOM-SUSTAINCOM 2016), 2016, : 242 - 249
  • [27] Improving Bregman k-means
    Ashour, Wesam
    Fyfe, Colin
    INTERNATIONAL JOURNAL OF DATA MINING MODELLING AND MANAGEMENT, 2014, 6 (01) : 65 - 82
  • [28] Comparative Study of K-Means, Pam and Rough K-Means Algorithms Using Cancer Datasets
    Kumar, Parvesh
    Wasan, Krishan
    COMPUTING, COMMUNICATION, AND CONTROL, 2011, 1 : 136 - 140
  • [29] An adaptive enhancement algorithm for infrared video based on modified k-means clustering
    Zhang, Linze
    Wang, Jingqi
    Wu, Wen
    INFRARED SENSORS, DEVICES, AND APPLICATIONS VI, 2016, 9974
  • [30] Global k-means plus plus : an effective relaxation of the global k-means clustering algorithm
    Vardakas, Georgios
    Likas, Aristidis
    APPLIED INTELLIGENCE, 2024, 54 (19) : 8876 - 8888