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
  • [1] Adaptive K-Means clustering algorithm
    Chen, Hailin
    Wu, Xiuqing
    Hu, Junhua
    MIPPR 2007: PATTERN RECOGNITION AND COMPUTER VISION, 2007, 6788
  • [2] Adapting k-means for graph clustering
    Sieranoja, Sami
    Franti, Pasi
    KNOWLEDGE AND INFORMATION SYSTEMS, 2022, 64 (01) : 115 - 142
  • [3] Adaptive Initialization Method for K-Means Algorithm
    Yang, Jie
    Wang, Yu-Kai
    Yao, Xin
    Lin, Chin-Teng
    FRONTIERS IN ARTIFICIAL INTELLIGENCE, 2021, 4
  • [4] Empirical Evaluation of K-Means, Bisecting K-Means, Fuzzy C-Means and Genetic K-Means Clustering Algorithms
    Banerjee, Shreya
    Choudhary, Ankit
    Pal, Somnath
    2015 IEEE INTERNATIONAL WIE CONFERENCE ON ELECTRICAL AND COMPUTER ENGINEERING (WIECON-ECE), 2015, : 172 - 176
  • [5] Deep k-Means: Jointly clustering with k-Means and learning representations
    Fard, Maziar Moradi
    Thonet, Thibaut
    Gaussier, Eric
    PATTERN RECOGNITION LETTERS, 2020, 138 : 185 - 192
  • [6] An Adaptive Algorithm of K-means on HSA Platform
    Bao, Zhenshan
    Luo, Qi
    Zhang, Wenbo
    PROCEEDINGS OF THE 2016 INTERNATIONAL CONFERENCE ON COMPUTER SCIENCE AND ELECTRONIC TECHNOLOGY, 2016, 48 : 136 - 139
  • [7] A Reconfigurable 64-Dimension K-Means Clustering Accelerator With Adaptive Overflow Control
    Du, Li
    Du, Yuan
    Chang, Mau-Chung Frank
    IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS, 2020, 67 (04) : 760 - 764
  • [8] k-means: A revisit
    Zhao, Wan-Lei
    Deng, Cheng-Hao
    Ngo, Chong-Wah
    NEUROCOMPUTING, 2018, 291 : 195 - 206
  • [9] Soil data clustering by using K-means and fuzzy K-means algorithm
    Hot, Elma
    Popovic-Bugarin, Vesna
    2015 23RD TELECOMMUNICATIONS FORUM TELFOR (TELFOR), 2015, : 890 - 893
  • [10] Graph k-Anonymity through k-Means and as Modular Decomposition
    Stokes, Klara
    SECURE IT SYSTEMS, NORDSEC 2013, 2013, 8208 : 263 - 278