{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,1]],"date-time":"2026-06-01T23:56:38Z","timestamp":1780358198044,"version":"3.54.1"},"publisher-location":"Cham","reference-count":24,"publisher":"Springer Nature Switzerland","isbn-type":[{"value":"9783032083296","type":"print"},{"value":"9783032083302","type":"electronic"}],"license":[{"start":{"date-parts":[[2025,10,14]],"date-time":"2025-10-14T00:00:00Z","timestamp":1760400000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,10,14]],"date-time":"2025-10-14T00:00:00Z","timestamp":1760400000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026]]},"abstract":"<jats:title>Abstract<\/jats:title>\n          <jats:p>As machine learning models become increasingly prevalent in time series applications, Explainable Artificial Intelligence (XAI) methods are essential for understanding their predictions. Within XAI, feature attribution methods aim to identify which input features contribute the most to a model\u2019s prediction, with their evaluation typically relying on perturbation-based metrics. Through systematic empirical analysis across multiple datasets, model architectures, and perturbation strategies, we reveal previously overlooked class-dependent effects in these metrics: they show varying effectiveness across classes, achieving strong results for some while remaining less sensitive to others. In particular, we find that the most effective perturbation strategies often demonstrate the most pronounced class differences. Our analysis suggests that these effects arise from the learned biases of classifiers, indicating that perturbation-based evaluation may reflect specific model behaviors rather than intrinsic attribution quality. We propose an evaluation framework with a class-aware penalty term to help assess and account for these effects in evaluating feature attributions, offering particular value for class-imbalanced datasets. Although our analysis focuses on time series classification, these class-dependent effects likely extend to other structured data domains where perturbation-based evaluation is common (Code and results are available at <jats:ext-link xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" xlink:href=\"https:\/\/github.com\/gregorbaer\/class-perturbation-effects\" ext-link-type=\"uri\">https:\/\/github.com\/gregorbaer\/class-perturbation-effects<\/jats:ext-link>.).<\/jats:p>","DOI":"10.1007\/978-3-032-08330-2_14","type":"book-chapter","created":{"date-parts":[[2025,10,13]],"date-time":"2025-10-13T03:10:48Z","timestamp":1760325048000},"page":"292-314","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["Class-Dependent Perturbation Effects in\u00a0Evaluating Time Series Attributions"],"prefix":"10.1007","author":[{"ORCID":"https:\/\/orcid.org\/0009-0002-9918-1376","authenticated-orcid":false,"given":"Gregor","family":"Baer","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8035-2887","authenticated-orcid":false,"given":"Isel","family":"Grau","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9811-1881","authenticated-orcid":false,"given":"Chao","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5197-3986","authenticated-orcid":false,"given":"Pieter","family":"Van Gorp","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2025,10,14]]},"reference":[{"key":"14_CR1","doi-asserted-by":"publisher","first-page":"1293","DOI":"10.1109\/JAS.2019.1911747","volume":"6","author":"HA Dau","year":"2019","unstructured":"Dau, H.A., et al.: The UCR time series archive. IEEE\/CAA J. Automatica Sinica 6, 1293\u20131305 (2019). https:\/\/doi.org\/10.1109\/JAS.2019.1911747","journal-title":"IEEE\/CAA J. Automatica Sinica"},{"key":"14_CR2","doi-asserted-by":"publisher","unstructured":"Doshi-Velez, F., Kim, B.: Considerations for evaluation and generalization in interpretable machine learning. In: Escalante, H.J., Escalera, S., Guyon, I., Bar\u00f3, X., G\u00fc\u00e7l\u00fct\u00fcrk, Y., G\u00fc\u00e7l\u00fc, U., van Gerven, M. (eds.) Explainable and Interpretable Models in Computer Vision and Machine Learning, pp. 3\u201317. Springer (2018). https:\/\/doi.org\/10.1007\/978-3-319-98131-4_1","DOI":"10.1007\/978-3-319-98131-4_1"},{"key":"14_CR3","doi-asserted-by":"publisher","unstructured":"Fawaz, H.I., Forestier, G., Weber, J., Idoumghar, L., Muller, P.A.: Deep learning for time series classification: a review. Data Mining Knowl. Disc. 33, 917\u2013963 (2019). https:\/\/doi.org\/10.1007\/s10618-019-00619-1","DOI":"10.1007\/s10618-019-00619-1"},{"key":"14_CR4","doi-asserted-by":"crossref","unstructured":"Fong, R.C., Vedaldi, A.: Interpretable explanations of black boxes by meaningful perturbation. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 3429\u20133437 (2017). https:\/\/openaccess.thecvf.com\/content_iccv_2017\/html\/Fong_Interpretable_Explanations_of_ICCV_2017_paper.html","DOI":"10.1109\/ICCV.2017.371"},{"key":"14_CR5","doi-asserted-by":"publisher","unstructured":"H\u00f6llig, J., Kulbach, C., Thoma, S.: TSInterpret: a python package for the interpretability of time series classification. J. Open Source Softw. 8, 5220 (2023). https:\/\/doi.org\/10.21105\/joss.05220","DOI":"10.21105\/joss.05220"},{"key":"14_CR6","doi-asserted-by":"publisher","unstructured":"Fawaz, I.H., et al.: InceptionTime: finding AlexNet for time series classification. Data Mining Knowl. Disc. 34, 1936\u20131962 (2020). https:\/\/doi.org\/10.1007\/s10618-020-00710-y","DOI":"10.1007\/s10618-020-00710-y"},{"key":"14_CR7","unstructured":"Lundberg, S.M., Lee, S.I.: A Unified approach to interpreting model predictions. In: Guyon, I., Luxburg, U.V., Bengio, S., Wallach, H., Fergus, R., Vishwanathan, S., Garnett, R. (eds.) Advances in Neural Information Processing Systems, vol.\u00a030 (2017). https:\/\/proceedings.neurips.cc\/paper_files\/paper\/2017\/file\/8a20a8621978632d76c43dfd28b67767-Paper.pdf"},{"key":"14_CR8","doi-asserted-by":"publisher","unstructured":"Mercier, D., Bhatt, J., Dengel, A., Ahmed, S.: Time to Focus: A Comprehensive Benchmark Using Time Series Attribution Methods (2022). https:\/\/doi.org\/10.48550\/arXiv.2202.03759","DOI":"10.48550\/arXiv.2202.03759"},{"key":"14_CR9","doi-asserted-by":"publisher","unstructured":"Nauta, M., et al.: From anecdotal evidence to quantitative evaluation methods: a systematic review on evaluating explainable AI. ACM Comput. Surv. 55, 1\u201342 (2023). https:\/\/doi.org\/10.1145\/3583558","DOI":"10.1145\/3583558"},{"key":"14_CR10","doi-asserted-by":"publisher","unstructured":"Nguyen, T.T., Le\u00a0Nguyen, T., Ifrim, G.: Robust explainer recommendation for time series classification. Data Mining Knowl. Disc. 38, 3372\u20133413 (2024). https:\/\/doi.org\/10.1007\/s10618-024-01045-8","DOI":"10.1007\/s10618-024-01045-8"},{"issue":"4","key":"14_CR11","doi-asserted-by":"publisher","first-page":"2104","DOI":"10.1109\/TPAMI.2023.3331846","volume":"46","author":"Y Rong","year":"2024","unstructured":"Rong, Y., et al.: Towards human-centered explainable AI: a survey of user studies for model explanations. IEEE Trans. Pattern Anal. Mach. Intell. 46(4), 2104\u20132122 (2024). https:\/\/doi.org\/10.1109\/TPAMI.2023.3331846","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"14_CR12","doi-asserted-by":"publisher","unstructured":"Samek, W., Binder, A., Montavon, G., Lapuschkin, S., M\u00fcller, K.R.: Evaluating the visualization of what a deep neural network has learned. IEEE Trans. Neural Netw. Learn. Syst. 28, 2660\u20132673 (2017). https:\/\/doi.org\/10.1109\/TNNLS.2016.2599820","DOI":"10.1109\/TNNLS.2016.2599820"},{"key":"14_CR13","doi-asserted-by":"publisher","unstructured":"Schlegel, U., Arnout, H., El-Assady, M., Oelke, D., Keim, D.A.: Towards a rigorous evaluation of XAI methods on time series. In: 2019 IEEE\/CVF International Conference on Computer Vision Workshop (ICCVW), pp. 4197\u20134201 (2019). https:\/\/doi.org\/10.1109\/ICCVW.2019.00516","DOI":"10.1109\/ICCVW.2019.00516"},{"key":"14_CR14","doi-asserted-by":"publisher","unstructured":"Schlegel, U., Keim, D.A.: A deep dive into perturbations as evaluation technique for time series XAI. In: Longo, L. (ed.) Explainable Artificial Intelligence, vol.\u00a01903, pp. 165\u2013180. Springer Nature Switzerland, Cham (2023). https:\/\/doi.org\/10.1007\/978-3-031-44070-0_9","DOI":"10.1007\/978-3-031-44070-0_9"},{"key":"14_CR15","doi-asserted-by":"publisher","unstructured":"Schlegel, U., Keim, D.A.: Introducing the attribution stability indicator: a measure for time series XAI attributions. In: ECML-PKDD Workshop XAI-TS: Explainable AI for Time Series: Advances and Applications (2023). https:\/\/doi.org\/10.48550\/arXiv.2310.04178","DOI":"10.48550\/arXiv.2310.04178"},{"key":"14_CR16","unstructured":"Schulz, K., Sixt, L., Tombari, F., Landgraf, T.: Restricting the flow: information bottlenecks for attribution. In: International Conference on Learning Representations (2020). https:\/\/openreview.net\/forum?id=S1xWh1rYwB"},{"key":"14_CR17","doi-asserted-by":"publisher","unstructured":"Serramazza, D.I., Nguyen, T.L., Ifrim, G.: Improving the\u00a0evaluation and\u00a0actionability of\u00a0explanation methods for\u00a0multivariate time series classification. In: Bifet, A., Davis, J., Krilavi\u010dius, T., Kull, M., Ntoutsi, E., \u017dliobait\u0117, I. (eds.) Machine Learning and Knowledge Discovery in Databases. Research Track, pp. 177\u2013195 (2024). https:\/\/doi.org\/10.1007\/978-3-031-70359-1_11","DOI":"10.1007\/978-3-031-70359-1_11"},{"key":"14_CR18","doi-asserted-by":"publisher","unstructured":"\u0160imi\u0107, I., Sabol, V., Veas, E.: Perturbation effect: a metric to counter misleading validation of feature attribution. In: Proceedings of the 31st ACM International Conference on Information and Knowledge Management, pp. 1798\u20131807 (2022). https:\/\/doi.org\/10.1145\/3511808.3557418","DOI":"10.1145\/3511808.3557418"},{"key":"14_CR19","doi-asserted-by":"publisher","unstructured":"Simonyan, K., Vedaldi, A., Zisserman, A.: Deep inside convolutional networks: visualising image classification models and saliency maps. In: Proceedings of the International Conference on Learning Representations (ICLR) (2014). https:\/\/doi.org\/10.48550\/arXiv.1312.6034","DOI":"10.48550\/arXiv.1312.6034"},{"key":"14_CR20","doi-asserted-by":"publisher","unstructured":"Smilkov, D., Thorat, N., Kim, B., Vi\u00e9gas, F., Wattenberg, M.: SmoothGrad: removing noise by adding noise (2017). https:\/\/doi.org\/10.48550\/arXiv.1706.03825","DOI":"10.48550\/arXiv.1706.03825"},{"key":"14_CR21","unstructured":"Sundararajan, M., Taly, A., Yan, Q.: Axiomatic attribution for deep networks. In: Proceedings of the 34th International Conference on Machine Learning, pp. 3319\u20133328 (2017). https:\/\/proceedings.mlr.press\/v70\/sundararajan17a.html"},{"key":"14_CR22","doi-asserted-by":"publisher","unstructured":"Theissler, A., Spinnato, F., Schlegel, U., Guidotti, R.: Explainable AI for time series classification: a review, taxonomy and research directions. IEEE Access 10, 100700\u2013100724 (2022). https:\/\/doi.org\/10.1109\/ACCESS.2022.3207765","DOI":"10.1109\/ACCESS.2022.3207765"},{"key":"14_CR23","doi-asserted-by":"publisher","unstructured":"Turb\u00e9, H., Bjelogrlic, M., Lovis, C., Mengaldo, G.: Evaluation of post-hoc interpretability methods in time-series classification. Nat. Mach. Intell. 5, 250\u2013260 (2023). https:\/\/doi.org\/10.1038\/s42256-023-00620-w","DOI":"10.1038\/s42256-023-00620-w"},{"key":"14_CR24","doi-asserted-by":"publisher","unstructured":"Wang, Z., Yan, W., Oates, T.: Time series classification from scratch with deep neural networks: a strong baseline. In: 2017 International Joint Conference on Neural Networks (IJCNN), pp. 1578\u20131585 (2017). https:\/\/doi.org\/10.1109\/IJCNN.2017.7966039","DOI":"10.1109\/IJCNN.2017.7966039"}],"container-title":["Communications in Computer and Information Science","Explainable Artificial Intelligence"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/978-3-032-08330-2_14","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,13]],"date-time":"2025-10-13T03:10:53Z","timestamp":1760325053000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/978-3-032-08330-2_14"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,14]]},"ISBN":["9783032083296","9783032083302"],"references-count":24,"URL":"https:\/\/doi.org\/10.1007\/978-3-032-08330-2_14","relation":{},"ISSN":["1865-0929","1865-0937"],"issn-type":[{"value":"1865-0929","type":"print"},{"value":"1865-0937","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,10,14]]},"assertion":[{"value":"14 October 2025","order":1,"name":"first_online","label":"First Online","group":{"name":"ChapterHistory","label":"Chapter History"}},{"value":"All authors declare that they have no conflicts of interest.","order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Disclosure of Interests"}},{"value":"xAI","order":1,"name":"conference_acronym","label":"Conference Acronym","group":{"name":"ConferenceInfo","label":"Conference Information"}},{"value":"World Conference on Explainable Artificial Intelligence","order":2,"name":"conference_name","label":"Conference Name","group":{"name":"ConferenceInfo","label":"Conference Information"}},{"value":"Istanbul","order":3,"name":"conference_city","label":"Conference City","group":{"name":"ConferenceInfo","label":"Conference Information"}},{"value":"T\u00fcrkiye","order":4,"name":"conference_country","label":"Conference Country","group":{"name":"ConferenceInfo","label":"Conference Information"}},{"value":"2025","order":5,"name":"conference_year","label":"Conference Year","group":{"name":"ConferenceInfo","label":"Conference Information"}},{"value":"9 July 2025","order":7,"name":"conference_start_date","label":"Conference Start Date","group":{"name":"ConferenceInfo","label":"Conference Information"}},{"value":"11 July 2025","order":8,"name":"conference_end_date","label":"Conference End Date","group":{"name":"ConferenceInfo","label":"Conference Information"}},{"value":"3","order":9,"name":"conference_number","label":"Conference Number","group":{"name":"ConferenceInfo","label":"Conference Information"}},{"value":"xai2025","order":10,"name":"conference_id","label":"Conference ID","group":{"name":"ConferenceInfo","label":"Conference Information"}},{"value":"https:\/\/xaiworldconference.com\/2025\/","order":11,"name":"conference_url","label":"Conference URL","group":{"name":"ConferenceInfo","label":"Conference Information"}}]}}