{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,3]],"date-time":"2026-11-03T11:05:43Z","timestamp":1793703943016,"version":"4.5.1"},"reference-count":37,"publisher":"American Chemical Society (ACS)","issue":"17","license":[{"start":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T00:00:00Z","timestamp":1786665600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T00:00:00Z","timestamp":1786665600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"},{"start":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T00:00:00Z","timestamp":1786665600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-045"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,9,14]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Selecting a single best machine-learning regression model from a set of competing models can be challenging. While models selected based on cross-validation performance do not guarantee good predictions on external data, models selected solely on external validation performance do not ascertain precise predictions for other external sets. Therefore, we propose three quantitative metrics to guide modelers in selecting the best model using the modeling set information only. Three quantitative data sets of varying sizes and complexities were considered. Each data set was randomly split into a modeling set and an independent test set. The modeling set was further split thrice to generate training and validation sets. Various machine-learning models were developed and validated against the validation sets. Our proposed metrics were computed for each model using only training and validation set performances. The novel metric values guided the selection of the best models, which also demonstrated expected performance on the independent test set. Furthermore, successful applications of our framework on two additional benchmark data sets demonstrated its wider generalizability.<\/jats:p>","DOI":"10.1021\/acs.jcim.6c01964","type":"journal-article","created":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T10:02:21Z","timestamp":1786701741000},"page":"10835-10846","source":"Crossref","is-referenced-by-count":1,"title":["<i>Q<\/i>\n                    performance\n          2: Toward Unbiased Selection of Machine-Learning\nRegression Models"],"prefix":"10.1021","volume":"66","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-8468-0784","authenticated-orcid":true,"given":"Arkaprava","family":"Banerjee","sequence":"first","affiliation":[{"name":"Jadavpur University , , ,","place":["Kolkata, India, 700032"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4486-8074","authenticated-orcid":true,"given":"Kunal","family":"Roy","sequence":"additional","affiliation":[{"name":"Jadavpur University , , ,","place":["Kolkata, India, 700032"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"316","published-online":{"date-parts":[[2026,8,14]]},"reference":[{"key":"2026091409084490800_cit1","doi-asserted-by":"publisher","first-page":"3703","DOI":"10.1007\/s11030-024-11056-8","article-title":"A bibliometric\nanalysis of the Cheminformatics\/QSAR\nliterature (2000\u20132023) for predictive modeling in data science\nusing the SCOPUS database","volume":"29","author":"Banerjee","year":"2025","journal-title":"Mol. Diversity"},{"key":"2026091409084490800_cit2","doi-asserted-by":"publisher","first-page":"3525","DOI":"10.1039\/D0CS00098A","article-title":"QSAR without borders","volume":"49","author":"Muratov","year":"2020","journal-title":"Chem. Soc. Rev."},{"key":"2026091409084490800_cit3","doi-asserted-by":"publisher","first-page":"1539","DOI":"10.1021\/acs.jcim.5c02639","article-title":"NeuroTDPi: Interpretable\ndeep learning models with multimodal fusion for identifying neurotoxic\ncompounds","volume":"66","author":"Liu","year":"2026","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409084490800_cit4","doi-asserted-by":"publisher","DOI":"10.1016\/j.jhazmat.2025.137575","article-title":"Web server-based deep\nlearning-driven predictive models for respiratory toxicity of environmental\nchemicals: Mechanistic insights and interpretability","volume":"489","author":"Li","year":"2025","journal-title":"J. Hazard. Mater."},{"key":"2026091409084490800_cit5","doi-asserted-by":"publisher","first-page":"95","DOI":"10.1002\/(SICI)1099-128X(199603)10:2%3C95:AID-CEM407%3E3.0.CO;2-M","article-title":"Single and\ndomain mode variable selection in 3D QSAR applications","volume":"10","author":"Norinder","year":"1996","journal-title":"J. Chemom."},{"key":"2026091409084490800_cit6","doi-asserted-by":"publisher","first-page":"2553","DOI":"10.1021\/jm970732a","article-title":"Three-dimensional quantitative similarity\u2013activity\nrelationships (3D QSiAR) from SEAL similarity matrices","volume":"41","author":"Kubinyi","year":"1998","journal-title":"J. Med. Chem."},{"key":"2026091409084490800_cit7","doi-asserted-by":"publisher","first-page":"269","DOI":"10.1016\/S1093-3263(01)00123-1","article-title":"Beware of q2!","volume":"20","author":"Golbraikh","year":"2002","journal-title":"J.\nMol. Graphics Modell."},{"key":"2026091409084490800_cit8","doi-asserted-by":"publisher","first-page":"1176","DOI":"10.1039\/D4DD00393D","article-title":"SurfPro \u2013 a curated\ndatabase and predictive model of experimental properties of surfactants","volume":"4","author":"Ho\u0308dl","year":"2025","journal-title":"Digital Discovery"},{"key":"2026091409084490800_cit9","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2024.108731","article-title":"DeepRA: A\nnovel deep learning-read-across framework and its application in non-sugar\nsweeteners mutagenicity prediction","volume":"178","author":"Srisongkram","year":"2024","journal-title":"Comput.\nBiol. Med."},{"key":"2026091409084490800_cit10","doi-asserted-by":"publisher","first-page":"2320","DOI":"10.1021\/ci200211n","article-title":"Real external predictivity of QSAR models: How to evaluate\nit? Comparison of different validation criteria and proposal of using\nthe concordance correlation coefficient","volume":"51","author":"Chirico","year":"2011","journal-title":"J.\nChem. Inf. Model."},{"key":"2026091409084490800_cit11","doi-asserted-by":"publisher","first-page":"837","DOI":"10.1007\/s10822-013-9680-4","article-title":"Estimation of influential points\nin any data set from coefficient of determination and its leave-one-out\ncross-validated counterpart","volume":"27","author":"To\u0301th","year":"2013","journal-title":"J. Comput.-Aided\nMol. Des."},{"key":"2026091409084490800_cit12","doi-asserted-by":"publisher","first-page":"36","DOI":"10.1021\/acs.jcim.6b00625","article-title":"In silico prediction\nof physicochemical properties of environmental chemicals using molecular\nfingerprints and machine learning","volume":"57","author":"Zang","year":"2017","journal-title":"J. Chem.\nInf. Model."},{"key":"2026091409084490800_cit13","doi-asserted-by":"publisher","first-page":"311","DOI":"10.1002\/minf.201400030","article-title":"External evaluation\nof QSAR models, in addition to cross-validation: verification of predictive\ncapability on totally new chemicals","volume":"33","author":"Gramatica","year":"2014","journal-title":"Mol. Inf."},{"key":"2026091409084490800_cit14","doi-asserted-by":"publisher","first-page":"476","DOI":"10.1002\/minf.201000061","article-title":"Best Practices\nfor QSAR Model Development, Validation, and Exploitation","volume":"29","author":"Tropsha","year":"2010","journal-title":"Mol. Inf."},{"key":"2026091409084490800_cit15","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-032-10081-8","volume-title":"Activity Cliffs: Where\nQSAR Predictions Fail","author":"Roy","year":"2025"},{"key":"2026091409084490800_cit16","doi-asserted-by":"crossref","unstructured":"Banerjee, A.; Roy, K.\n          MCCPerformance: Towards unbiased\nselection of machine learning classification models \u2013 Case\nstudies on environmental QSAR endpoints\n          ChemRxiv\n          2026\n          10.26434\/chemrxiv.15005752\/v1.","DOI":"10.26434\/chemrxiv.15005752\/v1"},{"key":"2026091409084490800_cit17","doi-asserted-by":"publisher","first-page":"2570","DOI":"10.1021\/ci300338w","article-title":"Does\nrational selection\nof training and test sets improve the outcome of QSAR modeling?","volume":"52","author":"Martin","year":"2012","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409084490800_cit18","doi-asserted-by":"publisher","DOI":"10.1007\/s10822-026-00854-x","article-title":"A round-robin exercise\nfor the precise prediction of aqueous solubility of organic chemicals\nusing chemometric, machine learning, and stacking ensemble of deep\nlearning models","volume":"40","author":"Banerjee","year":"2026","journal-title":"J. Comput.-Aided Mol. Des."},{"key":"2026091409084490800_cit19","doi-asserted-by":"publisher","first-page":"621","DOI":"10.1080\/1062936X.2025.2540820","article-title":"Machine learning-based\nq-RASAR modelling for the in silico identification\nof novel \u03b17nAChR agonists for anti-Alzheimer\u2019s drug discovery","volume":"36","author":"Kumar","year":"2025","journal-title":"SAR QSAR Environ. Res."},{"key":"2026091409084490800_cit20","doi-asserted-by":"publisher","DOI":"10.1016\/j.mtcomm.2024.110430","article-title":"Predictive cheminformatics modeling\nof reorganization energy (RE)\nfor p-type organic semiconductors: Integration of quantitative read-across\nstructure-property relationship (q-RASPR) and stacking regression\nanalysis","volume":"41","author":"Pandey","year":"2024","journal-title":"Mater. Today Commun."},{"key":"2026091409084490800_cit21","unstructured":"Landrum, G.\n          \n          RDKit Documentation, 2013. https:\/\/ftp.ccp4.ac.uk\/ccp4\/7.0\/unpacked\/checkout\/rdkit-Release_2015_03_1\/Docs\/Book\/RDKit.pdf."},{"key":"2026091409084490800_cit22","doi-asserted-by":"publisher","first-page":"147","DOI":"10.1016\/j.chemolab.2014.05.010","article-title":"A novel variable reduction\nmethod adapted from space-filling designs","volume":"136","author":"Ballabio","year":"2014","journal-title":"Chemom.\nIntell. Lab. Syst."},{"key":"2026091409084490800_cit23","volume-title":"Statistical Methods","author":"Snedecor","year":"1991","edition":"8"},{"key":"2026091409084490800_cit24","doi-asserted-by":"crossref","DOI":"10.1007\/978-1-4615-0377-4_5","article-title":"Artificial\nNeural Network","volume-title":"Interdisciplinary Computing\nin Java Programming","author":"Wang","year":"2003"},{"key":"2026091409084490800_cit25","unstructured":"Dorogush, A. V.; Ershov, V.; Gulin, A.\n          CatBoost: Gradient\nBoosting\nwith Categorical Features Support. 2018, arXiv:1810.11363. arXiv.org e-Print archive. https:\/\/arxiv.org\/abs\/1810.11363."},{"key":"2026091409084490800_cit26","unstructured":"Bentejac, C.; Csorgo, A.; Martinez-Munoz, G.\n          A Comparative\nAnalysis of XGBoost. 2019, arXiv:1911.01914.\narXiv.org e-Print archive. https:\/\/arxiv.org\/abs\/1911.01914."},{"key":"2026091409084490800_cit27","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1007\/s10994-006-6226-1","article-title":"Extremely randomized trees","volume":"63","author":"Geurts","year":"2006","journal-title":"Mach. Learn."},{"key":"2026091409084490800_cit28","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1023\/A:1010933404324","article-title":"Random forests","volume":"45","author":"Breiman","year":"2001","journal-title":"Mach. Learn."},{"key":"2026091409084490800_cit29","doi-asserted-by":"crossref","unstructured":"Divya, S. V.; Venkadesh, P.; Pavithra, R. J.; et al. In  An\nIntegrated XGBRF Framework for Early Identification of Parkinson\u2019s\nDisease, 2024 International Conference on Emerging Research\nin Computational Science (ICERCS); IEEE, 2024; pp 1\u20139.","DOI":"10.1109\/ICERCS63125.2024.10894797"},{"key":"2026091409084490800_cit30","doi-asserted-by":"publisher","first-page":"367","DOI":"10.1016\/S0167-9473(01)00065-2","article-title":"Stochastic\ngradient boosting","volume":"38","author":"Friedman","year":"2002","journal-title":"Comput. Stat. Data Anal."},{"key":"2026091409084490800_cit31","doi-asserted-by":"publisher","first-page":"273","DOI":"10.1007\/BF00994018","article-title":"Support-vector networks","volume":"20","author":"Cortes","year":"1995","journal-title":"Mach.\nLearn."},{"key":"2026091409084490800_cit32","doi-asserted-by":"publisher","first-page":"21","DOI":"10.1109\/TIT.1967.1053964","article-title":"Nearest neighbor pattern\nclassification","volume":"13","author":"Cover","year":"1967","journal-title":"IEEE\nTrans. Inf. Theory"},{"key":"2026091409084490800_cit33","doi-asserted-by":"publisher","first-page":"109","DOI":"10.1016\/S0169-7439(01)00155-1","article-title":"PLS-regression: a basic\ntool of chemometrics","volume":"58","author":"Wold","year":"2001","journal-title":"Chemom. Intell. Lab. Syst."},{"key":"2026091409084490800_cit34","doi-asserted-by":"publisher","first-page":"348","DOI":"10.1002\/1521-3838(200210)21:4%3C348:AID-QSAR348%3E3.0.CO;2-D","article-title":"From narcosis\nto hyperspace: The history of QSAR","volume":"21","author":"Kubinyi","year":"2002","journal-title":"Quant. Struc.\nAct. Rel."},{"key":"2026091409084490800_cit35","doi-asserted-by":"publisher","first-page":"694","DOI":"10.1002\/qsar.200610151","article-title":"Principles\nof QSAR models validation: internal and external","volume":"26","author":"Gramatica","year":"2007","journal-title":"QSAR Comb. Sci."},{"key":"2026091409084490800_cit36","doi-asserted-by":"publisher","first-page":"1316","DOI":"10.1021\/acs.jcim.5b00206","article-title":"Beware of R2: Simple, unambiguous\nassessment of the prediction accuracy of QSAR and QSPR models","volume":"55","author":"Alexander","year":"2015","journal-title":"J. Chem. Inf. Model."},{"key":"2026091409084490800_cit37","doi-asserted-by":"publisher","first-page":"516","DOI":"10.1002\/minf.201200134","article-title":"A large-scale\nempirical\nevaluation of cross-validation and external test set validation in\n(Q)SAR","volume":"32","author":"Gu\u0308tlein","year":"2013","journal-title":"Mol. Inf."}],"container-title":["Journal of Chemical Information\nand Modeling"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pubs.acs.org\/doi\/pdf\/10.1021\/acs.jcim.6c01964","content-type":"application\/pdf","content-version":"vor","intended-application":"unspecified"},{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article-pdf\/66\/17\/10835\/66542418\/acs.jcim.6c01964.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article-pdf\/66\/17\/10835\/66542418\/acs.jcim.6c01964.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,14]],"date-time":"2026-09-14T13:28:22Z","timestamp":1789392502000},"score":1,"resource":{"primary":{"URL":"https:\/\/pubs.acs.org\/jcisd8\/article\/66\/17\/10835\/5255235\/Qperformance2-Toward-Unbiased-Selection-of-Machine"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,14]]},"references-count":37,"journal-issue":{"issue":"17","published-online":{"date-parts":[[2026,8,14]]},"published-print":{"date-parts":[[2026,9,14]]}},"URL":"https:\/\/doi.org\/10.1021\/acs.jcim.6c01964","relation":{"has-preprint":[{"id-type":"doi","id":"10.26434\/chemrxiv.15004777\/v1","asserted-by":"object"}]},"ISSN":["1549-9596","1549-960X"],"issn-type":[{"value":"1549-9596","type":"print"},{"value":"1549-960X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,14]]}}}