Uncertainty-Aware PCA for Arbitrarily Distributed Data Modeled by Gaussian Mixture Models

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Klötzl, Daniel, Tastekin, Ozan, Hägele, David, Evers, Marina, Weiskopf, Daniel
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912823821467648
author Klötzl, Daniel
Tastekin, Ozan
Hägele, David
Evers, Marina
Weiskopf, Daniel
author_facet Klötzl, Daniel
Tastekin, Ozan
Hägele, David
Evers, Marina
Weiskopf, Daniel
contents Multidimensional data is often associated with uncertainties that are not well-described by normal distributions. In this work, we describe how such distributions can be projected to a low-dimensional space using uncertainty-aware principal component analysis (UAPCA). We propose to model multidimensional distributions using Gaussian mixture models (GMMs) and derive the projection from a general formulation that allows projecting arbitrary probability density functions. The low-dimensional projections of the densities exhibit more details about the distributions and represent them more faithfully compared to UAPCA mappings. Further, we support including user-defined weights between the different distributions, which allows for varying the importance of the multidimensional distributions. We evaluate our approach by comparing the distributions in low-dimensional space obtained by our method and UAPCA to those obtained by sample-based projections.
format Preprint
id arxiv_https___arxiv_org_abs_2508_13990
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Uncertainty-Aware PCA for Arbitrarily Distributed Data Modeled by Gaussian Mixture Models
Klötzl, Daniel
Tastekin, Ozan
Hägele, David
Evers, Marina
Weiskopf, Daniel
Machine Learning
Graphics
Multidimensional data is often associated with uncertainties that are not well-described by normal distributions. In this work, we describe how such distributions can be projected to a low-dimensional space using uncertainty-aware principal component analysis (UAPCA). We propose to model multidimensional distributions using Gaussian mixture models (GMMs) and derive the projection from a general formulation that allows projecting arbitrary probability density functions. The low-dimensional projections of the densities exhibit more details about the distributions and represent them more faithfully compared to UAPCA mappings. Further, we support including user-defined weights between the different distributions, which allows for varying the importance of the multidimensional distributions. We evaluate our approach by comparing the distributions in low-dimensional space obtained by our method and UAPCA to those obtained by sample-based projections.
title Uncertainty-Aware PCA for Arbitrarily Distributed Data Modeled by Gaussian Mixture Models
topic Machine Learning
Graphics
url https://arxiv.org/abs/2508.13990