A comprehensive analysis of perturbation methods in explainable AI feature attribution validation for neural time series classifiers

Abstract In domains where AI model predictions have significant consequences, such as industry, medicine, and finance, the need for explainable AI (XAI) is of utmost importance. However, ensuring that explanation methods provide faithful and trustworthy explanations requires rigorous validation. Fea...

Full description

Saved in:
Bibliographic Details
Main Authors: Ilija Šimić, Eduardo Veas, Vedran Sabol
Format: Article
Language:English
Published: Nature Portfolio 2025-07-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-09538-2
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849237725304061952
author Ilija Šimić
Eduardo Veas
Vedran Sabol
author_facet Ilija Šimić
Eduardo Veas
Vedran Sabol
author_sort Ilija Šimić
collection DOAJ
description Abstract In domains where AI model predictions have significant consequences, such as industry, medicine, and finance, the need for explainable AI (XAI) is of utmost importance. However, ensuring that explanation methods provide faithful and trustworthy explanations requires rigorous validation. Feature attribution methods (AMs) are among the most prevalent explanation methods, as they identify decisive aspects that influence model predictions through feature importance estimates. Evaluating the correctness of AMs is typically done by systematically perturbing features according to their estimated importance and measuring the impact on the classifier’s performance. This paper extends our previous work which revealed flaws in the most commonly used metric for validating AMs when applied to time series data. In this work we introduce a novel metric, the Consistency-Magnitude-Index, which facilitates a faithful assessment of feature importance attribution. Additionally, we introduce an adapted methodology for robust faithfulness evaluation, leveraging a set of diverse perturbation methods. Our work includes an extended evaluation of AMs on time series data, that presents the influence and importance of perturbation methods and region size selection in relation to dataset and model characteristics. Based on the results of our extensive evaluation, we provide guidelines for future AM faithfulness assessments. Finally, we demonstrate our methodology through a concrete multivariate time series example.
format Article
id doaj-art-e304e9235e2b433ba0a6c4af4269c74b
institution Kabale University
issn 2045-2322
language English
publishDate 2025-07-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-e304e9235e2b433ba0a6c4af4269c74b2025-08-20T04:01:52ZengNature PortfolioScientific Reports2045-23222025-07-0115113010.1038/s41598-025-09538-2A comprehensive analysis of perturbation methods in explainable AI feature attribution validation for neural time series classifiersIlija Šimić0Eduardo Veas1Vedran Sabol2Graz University of TechnologyGraz University of TechnologyKnow Center Research GmbHAbstract In domains where AI model predictions have significant consequences, such as industry, medicine, and finance, the need for explainable AI (XAI) is of utmost importance. However, ensuring that explanation methods provide faithful and trustworthy explanations requires rigorous validation. Feature attribution methods (AMs) are among the most prevalent explanation methods, as they identify decisive aspects that influence model predictions through feature importance estimates. Evaluating the correctness of AMs is typically done by systematically perturbing features according to their estimated importance and measuring the impact on the classifier’s performance. This paper extends our previous work which revealed flaws in the most commonly used metric for validating AMs when applied to time series data. In this work we introduce a novel metric, the Consistency-Magnitude-Index, which facilitates a faithful assessment of feature importance attribution. Additionally, we introduce an adapted methodology for robust faithfulness evaluation, leveraging a set of diverse perturbation methods. Our work includes an extended evaluation of AMs on time series data, that presents the influence and importance of perturbation methods and region size selection in relation to dataset and model characteristics. Based on the results of our extensive evaluation, we provide guidelines for future AM faithfulness assessments. Finally, we demonstrate our methodology through a concrete multivariate time series example.https://doi.org/10.1038/s41598-025-09538-2Explainable AIDeep learningAttribution methodsEvaluationTime seriesDDS
spellingShingle Ilija Šimić
Eduardo Veas
Vedran Sabol
A comprehensive analysis of perturbation methods in explainable AI feature attribution validation for neural time series classifiers
Scientific Reports
Explainable AI
Deep learning
Attribution methods
Evaluation
Time series
DDS
title A comprehensive analysis of perturbation methods in explainable AI feature attribution validation for neural time series classifiers
title_full A comprehensive analysis of perturbation methods in explainable AI feature attribution validation for neural time series classifiers
title_fullStr A comprehensive analysis of perturbation methods in explainable AI feature attribution validation for neural time series classifiers
title_full_unstemmed A comprehensive analysis of perturbation methods in explainable AI feature attribution validation for neural time series classifiers
title_short A comprehensive analysis of perturbation methods in explainable AI feature attribution validation for neural time series classifiers
title_sort comprehensive analysis of perturbation methods in explainable ai feature attribution validation for neural time series classifiers
topic Explainable AI
Deep learning
Attribution methods
Evaluation
Time series
DDS
url https://doi.org/10.1038/s41598-025-09538-2
work_keys_str_mv AT ilijasimic acomprehensiveanalysisofperturbationmethodsinexplainableaifeatureattributionvalidationforneuraltimeseriesclassifiers
AT eduardoveas acomprehensiveanalysisofperturbationmethodsinexplainableaifeatureattributionvalidationforneuraltimeseriesclassifiers
AT vedransabol acomprehensiveanalysisofperturbationmethodsinexplainableaifeatureattributionvalidationforneuraltimeseriesclassifiers
AT ilijasimic comprehensiveanalysisofperturbationmethodsinexplainableaifeatureattributionvalidationforneuraltimeseriesclassifiers
AT eduardoveas comprehensiveanalysisofperturbationmethodsinexplainableaifeatureattributionvalidationforneuraltimeseriesclassifiers
AT vedransabol comprehensiveanalysisofperturbationmethodsinexplainableaifeatureattributionvalidationforneuraltimeseriesclassifiers