{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,12]],"date-time":"2026-11-12T12:12:47Z","timestamp":1794485567486,"version":"4.9.1"},"reference-count":44,"publisher":"American Chemical Society (ACS)","issue":"14","license":[{"start":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T00:00:00Z","timestamp":1784160000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T00:00:00Z","timestamp":1784160000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"},{"start":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T00:00:00Z","timestamp":1784160000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-045"}],"funder":[{"DOI":"10.13039\/501100005089","name":"Natural Science Foundation of Beijing Municipality","doi-asserted-by":"publisher","award":["QY26313"],"award-info":[{"award-number":["QY26313"]}],"id":[{"id":"10.13039\/501100005089","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["22203008"],"award-info":[{"award-number":["22203008"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,7,27]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>While feature importance analysis in chemical and materials machine learning can often be sensitive to both the predictive model and the attribution rule, the robustness of these rankings is rarely quantified before they are used to gain physical insights. Here, we compare 26 feature importance pipelines spanning data-driven, model-based, and formula-based analyses on a metal\u2013support interaction data set anchored by an explicit SISSO equation, and we examine whether the same qualitative behavior recurs in high-entropy-alloy and halide perovskite data sets. Across the three benchmarks, we observe high intrafamily agreement but substantial interfamily variance. While a small subset of features remains stable across multiple families, several midranked features are highly family dependent, with their apparent importance shifting according to the underlying modeling assumptions. To ensure robust interpretability, we recommend that feature importance be reported by method family or correlation-based clusters, supplemented by resampling intervals.<\/jats:p>","DOI":"10.1021\/acs.jcim.6c00893","type":"journal-article","created":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T18:03:01Z","timestamp":1784224981000},"page":"7996-8007","source":"Crossref","is-referenced-by-count":2,"title":["Familywise Feature\nImportance Stability in Chemical\nand Materials Machine Learning"],"prefix":"10.1021","volume":"66","author":[{"given":"Ruilin","family":"Lai","sequence":"first","affiliation":[{"name":"Beijing Information Science and Technology University , , ,","place":["Beijing, P. R. China, 100101"]},{"name":"Beijing Information Science and Technology University , , ,","place":["Beijing, P. R. China, 100101"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7755-2530","authenticated-orcid":true,"given":"Xiaotong","family":"Liu","sequence":"additional","affiliation":[{"name":"Beijing Information Science and Technology University , , ,","place":["Beijing, P. R. China, 100101"]},{"name":"Beijing Information Science and Technology University , , ,","place":["Beijing, P. R. China, 100101"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yuhang","family":"Wang","sequence":"additional","affiliation":[{"name":"Beijing Information Science and Technology University , , ,","place":["Beijing, P. R. China, 100101"]},{"name":"Beijing Information Science and Technology University , , ,","place":["Beijing, P. R. China, 100101"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1422-1205","authenticated-orcid":true,"given":"Gian-Marco","family":"Rignanese","sequence":"additional","affiliation":[{"name":"UCLouvain , Chemin des E\u0301toiles 8 , ,","place":["Louvain-la-Neuve, Belgium, 1348"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5043-3088","authenticated-orcid":true,"given":"Wei-Xue","family":"Li","sequence":"additional","affiliation":[{"name":"University of Science and Technology of China , , ,","place":["Anhui, P. R. China, 230026"]},{"name":"University of Science and Technology of China , , ,","place":["Anhui, P. R. China, 230026"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"316","published-online":{"date-parts":[[2026,7,16]]},"reference":[{"key":"2026072709090207400_cit1","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-024-52519-8","article-title":"Machine learning-assisted\ndual-atom sites design with interpretable descriptors unifying electrocatalytic\nreactions","volume":"15","author":"Lin","year":"2024","journal-title":"Nat. Commun."},{"key":"2026072709090207400_cit2","doi-asserted-by":"publisher","first-page":"601","DOI":"10.1038\/s41570-025-00740-4","article-title":"Developing machine learning for heterogeneous catalysis\nwith experimental and computational data","volume":"9","author":"Bozal-Ginesta","year":"2025","journal-title":"Nat.\nRev. Chem."},{"key":"2026072709090207400_cit3","doi-asserted-by":"publisher","first-page":"121740","DOI":"10.1016\/j.ces.2025.121740","article-title":"Big data-driven\nmachine learning\ntransformation for atomic-scale heterogeneous catalyst design: a critical\nreview","volume":"313","author":"Zheng","year":"2025","journal-title":"Chem. Eng. Sci."},{"key":"2026072709090207400_cit4","doi-asserted-by":"publisher","first-page":"e01139","DOI":"10.1016\/j.cocom.2025.e01139","article-title":"Machine learning-driven materials\ndiscovery: Unlocking\nnext-generation functional materials - A review","volume":"45","author":"Nematov","year":"2025","journal-title":"Comput. Condens. Matter"},{"key":"2026072709090207400_cit5","doi-asserted-by":"publisher","first-page":"22","DOI":"10.1021\/accountsmr.3c00131","article-title":"Interpretable machine\nlearning for catalytic materials design toward sustainability","volume":"5","author":"Xin","year":"2024","journal-title":"Acc. Mater. Res."},{"key":"2026072709090207400_cit6","doi-asserted-by":"publisher","first-page":"175","DOI":"10.1038\/s41929-022-00744-z","article-title":"Interpretable machine\nlearning for\nknowledge generation in heterogeneous catalysis","volume":"5","author":"Esterhuizen","year":"2022","journal-title":"Nat. Catal."},{"key":"2026072709090207400_cit7","doi-asserted-by":"publisher","first-page":"11390","DOI":"10.1039\/D4CS00844H","article-title":"Unlocking the potential: machine\nlearning applications in electrocatalyst\ndesign for electrochemical hydrogen energy transformation","volume":"53","author":"Ding","year":"2024","journal-title":"Chem. Soc. Rev."},{"key":"2026072709090207400_cit8","doi-asserted-by":"publisher","first-page":"pgae339","DOI":"10.1093\/pnasnexus\/pgae339","article-title":"Predicting and understanding photocatalytic\nCO2 reduction reaction with IR spectroscopy-based interpretable machine\nlearning framework","volume":"3","author":"Wang","year":"2024","journal-title":"PNAS Nexus"},{"key":"2026072709090207400_cit9","doi-asserted-by":"publisher","first-page":"7428","DOI":"10.1021\/acscatal.3c00611","article-title":"Interpretable catalysis models using\nmachine learning with spectroscopic\ndescriptors","volume":"13","author":"Wang","year":"2023","journal-title":"ACS Catal."},{"key":"2026072709090207400_cit10","doi-asserted-by":"publisher","first-page":"122593","DOI":"10.1016\/j.apcatb.2023.122593","article-title":"Impacts of\ncatalyst and process parameters on Ni-catalyzed methane dry reforming\nvia interpretable machine learning","volume":"330","author":"Vellayappan","year":"2023","journal-title":"Appl. Catal.,\nB"},{"key":"2026072709090207400_cit11","doi-asserted-by":"publisher","first-page":"2301020","DOI":"10.1002\/advs.202301020","article-title":"Machine learning descriptors\nfor data-driven catalysis study","volume":"10","author":"Mou","year":"2023","journal-title":"Adv. Sci."},{"key":"2026072709090207400_cit12","doi-asserted-by":"publisher","first-page":"2708","DOI":"10.1039\/D3TA06316J","article-title":"Local descriptors-based machine learning\nmodel refined by cluster analysis for accurately predicting adsorption\nenergies on bimetallic alloys","volume":"12","author":"Usuga","year":"2024","journal-title":"J. Mater. Chem.\nA"},{"key":"2026072709090207400_cit13","doi-asserted-by":"publisher","first-page":"284","DOI":"10.1016\/S1872-2067(23)64467-5","article-title":"Fine-structure sensitive\ndeep learning\nframework for predicting catalytic properties with high precision","volume":"50","author":"Chen","year":"2023","journal-title":"Chin. J. Catal."},{"key":"2026072709090207400_cit14","doi-asserted-by":"publisher","first-page":"3836","DOI":"10.1039\/D2CY00284A","article-title":"Enhanced descriptor identification\nand mechanism understanding for\ncatalytic activity using a data-driven framework: revealing the importance\nof interactions between elementary steps","volume":"12","author":"Liao","year":"2022","journal-title":"Catal.\nSci. Technol."},{"key":"2026072709090207400_cit15","doi-asserted-by":"publisher","first-page":"e202400124","DOI":"10.1002\/open.202400124","article-title":"Atomic\nAdsorption Energies Prediction on Bimetallic Transition Metal Surfaces\nUsing an Interpretable Machine Learning-Accelerated Density Functional\nTheory Approach","volume":"14","author":"Tomacruz","year":"2025","journal-title":"ChemistryOpen"},{"key":"2026072709090207400_cit16","doi-asserted-by":"publisher","first-page":"21349","DOI":"10.1021\/acs.jpcc.4c05332","article-title":"A Machine learning and explainable AI framework tailored\nfor unbalanced experimental catalyst discovery","volume":"128","author":"Semnani","year":"2024","journal-title":"J. Phys. Chem. C"},{"key":"2026072709090207400_cit17","unstructured":"Vinchurkar, T.; Ock, J.; Farimani, A. B.\n          Explainable\nData-driven Modeling of Adsorption Energy in Heterogeneous Catalysis, 2024. https:\/\/arxiv.org\/abs\/2405.20397."},{"key":"2026072709090207400_cit18","doi-asserted-by":"publisher","DOI":"10.3390\/catal14030195","article-title":"Catalytic\nActivity Prediction of \u03b1-Diimino Nickel Precatalysts\ntoward Ethylene Polymerization by Machine Learning","volume":"14","author":"Abbas","year":"2024","journal-title":"Catalysts"},{"key":"2026072709090207400_cit19","doi-asserted-by":"publisher","first-page":"411","DOI":"10.1039\/D4DD00216D","article-title":"Catalytic resonance\ntheory: forecasting the flow of\nprogrammable catalytic loops","volume":"4","author":"Murphy","year":"2025","journal-title":"Digit. Discovery"},{"key":"2026072709090207400_cit20","unstructured":"Xu, G.; Duong, T. D.; Li, Q.; Liu, S.; Wang, X.\n          Causality Learning: ANew Perspective for Interpretable\nMachine Learning, 2021. https:\/\/arxiv.org\/abs\/2006.16789."},{"key":"2026072709090207400_cit21","doi-asserted-by":"publisher","first-page":"18","DOI":"10.1145\/3400051.3400058","article-title":"Causal Interpretability\nfor Machine Learning - Problems, Methods and Evaluation","volume":"22","author":"Moraffah","year":"2020","journal-title":"SIGKDD Explor. Newsl."},{"key":"2026072709090207400_cit22","doi-asserted-by":"crossref","unstructured":"Krishna, S.; Han, T.; Gu, A.; Wu, S.; Jabbari, S.; Lakkaraju, H.\n          The disagreement problem in explainable machine learning:\nA practitioner\u2019s perspective, 2022.","DOI":"10.21203\/rs.3.rs-2963888\/v1"},{"key":"2026072709090207400_cit23","doi-asserted-by":"publisher","first-page":"82","DOI":"10.1007\/s11222-021-10057-z","article-title":"Unrestricted\npermutation forces extrapolation:\nvariable importance requires at least one more model, or there is\nno free variable importance","volume":"31","author":"Hooker","year":"2021","journal-title":"Stat. Comput."},{"key":"2026072709090207400_cit24","doi-asserted-by":"publisher","first-page":"307","DOI":"10.1186\/1471-2105-9-307","article-title":"Conditional\nvariable importance for random forests","volume":"9","author":"Strobl","year":"2008","journal-title":"BMC\nBioinf."},{"key":"2026072709090207400_cit25","doi-asserted-by":"publisher","first-page":"2903","DOI":"10.1007\/s10618-022-00901-9","article-title":"Model-agnostic\nfeature importance and effects with dependent features: a conditional\nsubgroup approach","volume":"38","author":"Molnar","year":"2024","journal-title":"Data Min. Knowl. Discovery"},{"key":"2026072709090207400_cit26","doi-asserted-by":"publisher","first-page":"103502","DOI":"10.1016\/j.artint.2021.103502","article-title":"Explaining\nindividual predictions\nwhen features are dependent: More accurate approximations to Shapley\nvalues","volume":"298","author":"Aas","year":"2021","journal-title":"Artif. Intell."},{"key":"2026072709090207400_cit27","doi-asserted-by":"publisher","first-page":"083802","DOI":"10.1103\/PhysRevMaterials.2.083802","article-title":"SISSO: A compressed-sensing method for identifying\nthe best low-dimensional descriptor in an immensity of offered candidates","volume":"2","author":"Ouyang","year":"2018","journal-title":"Phys. Rev. Mater."},{"key":"2026072709090207400_cit28","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-021-22048-9","article-title":"Single-atom alloy catalysts designed by first-principles\ncalculations and artificial intelligence","volume":"12","author":"Han","year":"2021","journal-title":"Nat.\nCommun."},{"key":"2026072709090207400_cit29","doi-asserted-by":"publisher","first-page":"150","DOI":"10.1038\/s41524-025-01596-4","article-title":"Materials-discovery\nworkflow guided\nby symbolic regression for identifying acid-stable oxides for electrocatalysis","volume":"11","author":"Nair","year":"2025","journal-title":"npj Comput. Mater."},{"key":"2026072709090207400_cit30","doi-asserted-by":"publisher","first-page":"1016","DOI":"10.1557\/s43577-021-00165-6","article-title":"Materials genes of heterogeneous catalysis\nfrom clean experiments and artificial intelligence","volume":"46","author":"Foppa","year":"2021","journal-title":"MRS Bull."},{"key":"2026072709090207400_cit31","doi-asserted-by":"publisher","first-page":"112","DOI":"10.1038\/s41524-023-01063-y","article-title":"Accelerating materials-space\nexploration for thermal insulators by mapping materials properties\nvia artificial intelligence","volume":"9","author":"Purcell","year":"2023","journal-title":"npj Comput. Mater."},{"key":"2026072709090207400_cit32","doi-asserted-by":"publisher","first-page":"2","DOI":"10.1007\/s10462-023-10622-0","article-title":"Interpretable scientific discovery\nwith symbolic regression:\na review","volume":"57","author":"Makke","year":"2024","journal-title":"Artif. Intell. Rev."},{"key":"2026072709090207400_cit33","doi-asserted-by":"publisher","first-page":"915","DOI":"10.1126\/science.adp6034","article-title":"Nature of metal-support\ninteraction for metal catalysts on oxide\nsupports","volume":"386","author":"Wang","year":"2024","journal-title":"Science"},{"key":"2026072709090207400_cit34","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-025-24781-3","article-title":"Machine learning\nmodels\nwith SHAP for performance prediction of eco-friendly fiber reinforced\nmortars with glass waste","volume":"15","author":"Rahmani","year":"2025","journal-title":"Sci. Rep."},{"key":"2026072709090207400_cit35","doi-asserted-by":"publisher","first-page":"44","DOI":"10.1186\/s40537-024-00905-w","article-title":"Feature\nselection strategies: a comparative analysis of SHAP-value and importance-based\nmethods","volume":"11","author":"Wang","year":"2024","journal-title":"J. Big Data"},{"key":"2026072709090207400_cit36","doi-asserted-by":"publisher","first-page":"106910","DOI":"10.1016\/j.mtcomm.2023.106910","article-title":"MIC-SHAP: An ensemble\nfeature selection method for materials machine learning","volume":"37","author":"Wang","year":"2023","journal-title":"Mater. Today Commun."},{"key":"2026072709090207400_cit37","doi-asserted-by":"publisher","first-page":"26","DOI":"10.57041\/pjosr.v3i1.946","article-title":"Xgboost\nand random forest algorithms: an in depth analysis","volume":"3","author":"Fatima","year":"2023","journal-title":"Pak. J. Sci. Res."},{"key":"2026072709090207400_cit38","doi-asserted-by":"crossref","unstructured":"Chen, T.; Guestrin, C. In  XGBoost: A Scalable\nTree Boosting System, Proceedings of the 22nd ACM SIGKDD\nInternational Conference on Knowledge Discovery and Data Mining, 2016.","DOI":"10.1145\/2939672.2939785"},{"key":"2026072709090207400_cit39","doi-asserted-by":"publisher","first-page":"56","DOI":"10.1016\/j.jclinepi.2020.03.002","article-title":"Logistic regression\nwas as good as machine learning for predicting major chronic diseases","volume":"122","author":"Nusinovici","year":"2020","journal-title":"J. Clin. Epidemiol."},{"key":"2026072709090207400_cit40","doi-asserted-by":"crossref","unstructured":"Verbeek, M.\n          \n          Using linear regression\nto establish empirical relationships.  IZA\nWorld of Labor, 2017. 10.15185\/izawol.336.","DOI":"10.15185\/izawol.336"},{"key":"2026072709090207400_cit41","doi-asserted-by":"publisher","first-page":"117290","DOI":"10.1016\/j.cma.2024.117290","article-title":"A comprehensive\nand FAIR comparison between MLP and KAN representations for differential\nequations and operator networks","volume":"431","author":"Shukla","year":"2024","journal-title":"Comput. Methods\nAppl. Mech. Eng."},{"key":"2026072709090207400_cit42","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-024-55243-x","article-title":"A comparative study of 11 non-linear regression models\nhighlighting autoencoder, DBN, and SVR, enhanced by SHAP importance\nanalysis in soybean branching prediction","volume":"14","author":"Zhou","year":"2024","journal-title":"Sci.\nRep."},{"key":"2026072709090207400_cit43","doi-asserted-by":"publisher","DOI":"10.3390\/math12243935","article-title":"Exploring\nKernel Machines and Support Vector Machines: Principles, Techniques,\nand Future Directions","volume":"12","author":"Du","year":"2024","journal-title":"Mathematics"},{"key":"2026072709090207400_cit44","unstructured":"Wang, Y.; Ni, X. S.\n          A XGBoost risk model via feature\nselection and Bayesian hyper-parameter optimization, 2019. https:\/\/arxiv.org\/abs\/1901.08433."}],"container-title":["Journal of Chemical Information\nand Modeling"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pubs.acs.org\/doi\/pdf\/10.1021\/acs.jcim.6c00893","content-type":"application\/pdf","content-version":"vor","intended-application":"unspecified"},{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article-pdf\/66\/14\/7996\/65981379\/acs.jcim.6c00893.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article-pdf\/66\/14\/7996\/65981379\/acs.jcim.6c00893.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,27]],"date-time":"2026-07-27T13:29:30Z","timestamp":1785158970000},"score":1,"resource":{"primary":{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article\/66\/14\/7996\/5212281\/Familywise-Feature-Importance-Stability-in"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,16]]},"references-count":44,"journal-issue":{"issue":"14","published-online":{"date-parts":[[2026,7,16]]},"published-print":{"date-parts":[[2026,7,27]]}},"URL":"https:\/\/doi.org\/10.1021\/acs.jcim.6c00893","relation":{"has-part":[{"id-type":"doi","id":"https:\/\/doi.org\/10.1021\/acs.jcim.6c00893.s001","asserted-by":"subject"}]},"ISSN":["1549-9596","1549-960X"],"issn-type":[{"value":"1549-9596","type":"print"},{"value":"1549-960X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,16]]}}}