Guardado en:
Detalles Bibliográficos
Autores principales: Bae, Jiyeon, Jeon, Hyeon, Seo, Jinwook
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2507.02225
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912816093462528
author Bae, Jiyeon
Jeon, Hyeon
Seo, Jinwook
author_facet Bae, Jiyeon
Jeon, Hyeon
Seo, Jinwook
contents Evaluating the accuracy of dimensionality reduction (DR) projections in preserving the structure of high-dimensional data is crucial for reliable visual analytics. Diverse evaluation metrics targeting different structural characteristics have thus been developed. However, evaluations of DR projections can become biased if highly correlated metrics--those measuring similar structural characteristics--are inadvertently selected, favoring DR techniques that emphasize those characteristics. To address this issue, we propose a novel workflow that reduces bias in the selection of evaluation metrics by clustering metrics based on their empirical correlations rather than on their intended design characteristics alone. Our workflow works by computing metric similarity using pairwise correlations, clustering metrics to minimize overlap, and selecting a representative metric from each cluster. Quantitative experiments demonstrate that our approach improves the stability of DR evaluation, which indicates that our workflow contributes to mitigating evaluation bias.
format Preprint
id arxiv_https___arxiv_org_abs_2507_02225
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Metric Design != Metric Behavior: Improving Metric Selection for the Unbiased Evaluation of Dimensionality Reduction
Bae, Jiyeon
Jeon, Hyeon
Seo, Jinwook
Machine Learning
Evaluating the accuracy of dimensionality reduction (DR) projections in preserving the structure of high-dimensional data is crucial for reliable visual analytics. Diverse evaluation metrics targeting different structural characteristics have thus been developed. However, evaluations of DR projections can become biased if highly correlated metrics--those measuring similar structural characteristics--are inadvertently selected, favoring DR techniques that emphasize those characteristics. To address this issue, we propose a novel workflow that reduces bias in the selection of evaluation metrics by clustering metrics based on their empirical correlations rather than on their intended design characteristics alone. Our workflow works by computing metric similarity using pairwise correlations, clustering metrics to minimize overlap, and selecting a representative metric from each cluster. Quantitative experiments demonstrate that our approach improves the stability of DR evaluation, which indicates that our workflow contributes to mitigating evaluation bias.
title Metric Design != Metric Behavior: Improving Metric Selection for the Unbiased Evaluation of Dimensionality Reduction
topic Machine Learning
url https://arxiv.org/abs/2507.02225