Graph laplacians and their convergence on random neighborhood graphs

被引:0
作者
Hein, Matthias
Audibert, Jean-Yves
von Luxburg, Ulrike
机构
[1] Max Planck Inst Biol Cybernet, Dept Scholkopf, D-72076 Tubingen, Germany
[2] Ecole Natl Ponts & Chaussees, F-77455 Marnee Vallee, France
关键词
graphs; graph Laplacians; semi-supervised learning; spectral clustering; dimensionality reduction;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Given a sample from a probability measure with support on a submanifold in Euclidean space one can construct a neighborhood graph which can be seen as an approximation of the submanifold. The graph Laplacian of such a graph is used in several machine learning methods like semi-supervised learning, dimensionality reduction and clustering. In this paper we determine the pointwise limit of three different graph Laplacians used in the literature as the sample size increases and the neighborhood size approaches zero. We show that for a uniform measure on the submanifold all graph Laplacians have the same limit up to constants. However in the case of a non-uniform measure on the submanifold only the so called random walk graph Laplacian converges to the weighted Laplace-Beltrami operator.
引用
收藏
页码:1325 / 1368
页数:44
相关论文
共 50 条
[21]   HARDY SPACES ASSOCIATED TO THE DISCRETE LAPLACIANS ON GRAPHS AND BOUNDEDNESS OF SINGULAR INTEGRALS [J].
The Anh Bui ;
Duong, Xuan Thinh .
TRANSACTIONS OF THE AMERICAN MATHEMATICAL SOCIETY, 2014, 366 (07) :3451-3485
[22]   Statistical Analysis of Random Objects Via Metric Measure Laplacians [J].
Mordant, Gilles ;
Munk, Axel .
SIAM JOURNAL ON MATHEMATICS OF DATA SCIENCE, 2023, 5 (02) :528-557
[23]   Graph Laplacians and discrete reproducing kernel Hilbert spaces from restrictions [J].
Jorgensen, Palle ;
Tian, Feng .
STOCHASTIC ANALYSIS AND APPLICATIONS, 2016, 34 (04) :722-747
[24]   Semantic Visualization with Neighborhood Graph Regularization [J].
Le, Tuan M. V. ;
Lauw, Hady W. .
JOURNAL OF ARTIFICIAL INTELLIGENCE RESEARCH, 2016, 55 :1091-1133
[25]   Compact and Scalable Graph Neighborhood Sketching [J].
Akiba, Takuya ;
Yano, Yosuke .
KDD'16: PROCEEDINGS OF THE 22ND ACM SIGKDD INTERNATIONAL CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, 2016, :685-694
[26]   Approximations for the von Neumann and Re′nyi entropies of graphs with circulant type Laplacians [J].
Bebiano, Natalia ;
da Providencia, Joao ;
Xu, Wei-Ru .
ELECTRONIC RESEARCH ARCHIVE, 2022, 30 (05) :1864-1880
[27]   Computing with graphs and graph transformations [J].
Blostein, D ;
Schürr, A .
SOFTWARE-PRACTICE & EXPERIENCE, 1999, 29 (03) :197-217
[28]   Semi-supervised Learning Based on Joint Diffusion of Graph Functions and Laplacians [J].
Kim, Kwang In .
COMPUTER VISION - ECCV 2016, PT V, 2016, 9909 :713-729
[29]   Neighborhood Regularized l1-Graph [J].
Yang, Yingzhen ;
Feng, Jiashi ;
Yu, Jiahui ;
Yang, Jianchao ;
Kohli, Pushmeet ;
Huang, Thomas S. .
CONFERENCE ON UNCERTAINTY IN ARTIFICIAL INTELLIGENCE (UAI2017), 2017,
[30]   Infinite Random Geometric Graphs [J].
Bonato, Anthony ;
Janssen, Jeannette .
ANNALS OF COMBINATORICS, 2011, 15 (04) :597-617