DCFO: Density-Based Counterfactuals for Outliers -- Additional Material

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Amico, Tommaso, Matthews, Pernille, Krieger, Lena, Zimek, Arthur, Assent, Ira
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917543594164224
author Amico, Tommaso
Matthews, Pernille
Krieger, Lena
Zimek, Arthur
Assent, Ira
author_facet Amico, Tommaso
Matthews, Pernille
Krieger, Lena
Zimek, Arthur
Assent, Ira
contents Outlier detection identifies data points that significantly deviate from the majority of the data distribution. Explaining outliers is crucial for understanding the underlying factors that contribute to their detection, validating their significance, and identifying potential biases or errors. Effective explanations provide actionable insights, facilitating preventive measures to avoid similar outliers in the future. Counterfactual explanations clarify why specific data points are classified as outliers by identifying minimal changes required to alter their prediction. Although valuable, most existing counterfactual explanation methods overlook the unique challenges posed by outlier detection, and fail to target classical, widely adopted outlier detection algorithms. Local Outlier Factor (LOF) is one the most popular unsupervised outlier detection methods, quantifying outlierness through relative local density. Despite LOF's widespread use across diverse applications, it lacks interpretability. To address this limitation, we introduce Density-based Counterfactuals for Outliers (DCFO), a novel method specifically designed to generate counterfactual explanations for LOF. DCFO partitions the data space into regions where LOF behaves smoothly, enabling efficient gradient-based optimisation. Extensive experimental validation on 50 OpenML datasets demonstrates that DCFO consistently outperforms benchmarked competitors, offering superior proximity and validity of generated counterfactuals.
format Preprint
id arxiv_https___arxiv_org_abs_2512_10659
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DCFO: Density-Based Counterfactuals for Outliers -- Additional Material
Amico, Tommaso
Matthews, Pernille
Krieger, Lena
Zimek, Arthur
Assent, Ira
Machine Learning
Outlier detection identifies data points that significantly deviate from the majority of the data distribution. Explaining outliers is crucial for understanding the underlying factors that contribute to their detection, validating their significance, and identifying potential biases or errors. Effective explanations provide actionable insights, facilitating preventive measures to avoid similar outliers in the future. Counterfactual explanations clarify why specific data points are classified as outliers by identifying minimal changes required to alter their prediction. Although valuable, most existing counterfactual explanation methods overlook the unique challenges posed by outlier detection, and fail to target classical, widely adopted outlier detection algorithms. Local Outlier Factor (LOF) is one the most popular unsupervised outlier detection methods, quantifying outlierness through relative local density. Despite LOF's widespread use across diverse applications, it lacks interpretability. To address this limitation, we introduce Density-based Counterfactuals for Outliers (DCFO), a novel method specifically designed to generate counterfactual explanations for LOF. DCFO partitions the data space into regions where LOF behaves smoothly, enabling efficient gradient-based optimisation. Extensive experimental validation on 50 OpenML datasets demonstrates that DCFO consistently outperforms benchmarked competitors, offering superior proximity and validity of generated counterfactuals.
title DCFO: Density-Based Counterfactuals for Outliers -- Additional Material
topic Machine Learning
url https://arxiv.org/abs/2512.10659