Estimating Graph Dimension with Cross-validated Eigenvalues

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Chen, Fan, Roch, Sebastien, Rohe, Karl, Yu, Shuqi
Format: Preprint
Veröffentlicht: 2021
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917164217270272
author Chen, Fan
Roch, Sebastien
Rohe, Karl
Yu, Shuqi
author_facet Chen, Fan
Roch, Sebastien
Rohe, Karl
Yu, Shuqi
contents In applied multivariate statistics, estimating the number of latent dimensions or the number of clusters, $k$, is a fundamental and recurring problem. We study a sequence of statistics called "cross-validated eigenvalues." Under a large class of random graph models, including both Poisson and Bernoulli edges, without parametric assumptions, we provide a $p$-value for each cross-validated eigenvalue. It tests the null hypothesis that the sample eigenvector is orthogonal to (i.e., uncorrelated with) the true latent dimensions. This approach naturally adapts to problems where some dimensions are not statistically detectable. In scenarios where all $k$ dimensions can be estimated, we show that our procedure consistently estimates $k$. In simulations and data example, the proposed estimator compares favorably to alternative approaches in both computational and statistical performance.
format Preprint
id arxiv_https___arxiv_org_abs_2108_03336
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle Estimating Graph Dimension with Cross-validated Eigenvalues
Chen, Fan
Roch, Sebastien
Rohe, Karl
Yu, Shuqi
Methodology
Machine Learning
Social and Information Networks
Statistics Theory
In applied multivariate statistics, estimating the number of latent dimensions or the number of clusters, $k$, is a fundamental and recurring problem. We study a sequence of statistics called "cross-validated eigenvalues." Under a large class of random graph models, including both Poisson and Bernoulli edges, without parametric assumptions, we provide a $p$-value for each cross-validated eigenvalue. It tests the null hypothesis that the sample eigenvector is orthogonal to (i.e., uncorrelated with) the true latent dimensions. This approach naturally adapts to problems where some dimensions are not statistically detectable. In scenarios where all $k$ dimensions can be estimated, we show that our procedure consistently estimates $k$. In simulations and data example, the proposed estimator compares favorably to alternative approaches in both computational and statistical performance.
title Estimating Graph Dimension with Cross-validated Eigenvalues
topic Methodology
Machine Learning
Social and Information Networks
Statistics Theory
url https://arxiv.org/abs/2108.03336