A simulation experiment on bootstrap inference for Self-Organizing Maps

EasyChair Preprint 5982

Abstract

A stochastic simulation study is carried out to learn about quantification of uncertainties for Self-Organizing Maps (SOM). A mixture of Gaussian distributions is assumed as data generating process and the Monte Carlo generated samples are transformed according to the Kohonen (SOM) algorithm. Additionally, the original data matrix is resampled for a bootstrap quantification of parameter estimation uncertainties.

Keyphrases: Bootstrap, Self-Organizing Maps, stochastic simulation, uncertainty

@booklet{EasyChair:5982,
  author    = {José Manuel Mira and María Muñoz-Fernández de Córdoba and Gabriel Valverde and Beatriz González-Pérez},
  title     = {A simulation experiment on bootstrap inference for Self-Organizing Maps},
  howpublished = {EasyChair Preprint 5982},
  year      = {EasyChair, 2021}}