{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,1,31]],"date-time":"2026-01-31T02:02:07Z","timestamp":1769824927345,"version":"3.49.0"},"reference-count":29,"publisher":"Public Library of Science (PLoS)","issue":"5","license":[{"start":{"date-parts":[[2025,5,12]],"date-time":"2025-05-12T00:00:00Z","timestamp":1747008000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Open Access Publication Fund of the University of Duisburg-Essen"}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:sec id=\"sec001\">\n<jats:title>Background<\/jats:title>\n<jats:p>Many datasets in medicine and other branches of science are incomplete. In this article we compare various imputation algorithms for missing data.<\/jats:p>\n<\/jats:sec>\n<jats:sec id=\"sec002\">\n<jats:title>Objectives<\/jats:title>\n<jats:p>We take the point of view that it has already been decided that the imputation should be carried out using multiple imputation by chained equation and the only decision left is that of a subroutine for the one-dimensional imputations. The subroutines to be compared are predictive mean matching, weighted predictive mean matching, sampling, classification or regression trees and random forests.<\/jats:p>\n<\/jats:sec>\n<jats:sec id=\"sec003\">\n<jats:title>Methods<\/jats:title>\n<jats:p>We compare these subroutines on real data and on simulated data. We consider the estimation of expected values, variances and coefficients of linear regression models, logistic regression models and Cox regression models. As real data we use data of the survival times after the diagnosis of an obstructive coronary artery disease with systolic blood pressure, LDL, diabetes, smoking behavior and family history of premature heart diseases as variables for which values have to be imputed. While we are mainly interested in statistical properties like biases, mean squared errors or coverage probabilities of confidence intervals, we also have an eye on the computation time.<\/jats:p>\n<\/jats:sec>\n<jats:sec id=\"sec004\">\n<jats:title>Results<\/jats:title>\n<jats:p>Weighted predictive mean matching had to be excluded from the statistical comparison due to its enormous computation time. Among the remaining algorithms, in most situations we tested, predictive mean matching performed best.<\/jats:p>\n<\/jats:sec>\n<jats:sec id=\"sec005\">\n<jats:title>Novelty<\/jats:title>\n<jats:p>This is by far the largest comparison study for subroutines of multiple imputation by chained equations that has been performed up to now.<\/jats:p>\n<\/jats:sec>","DOI":"10.1371\/journal.pone.0319784","type":"journal-article","created":{"date-parts":[[2025,5,12]],"date-time":"2025-05-12T18:08:46Z","timestamp":1747073326000},"page":"e0319784","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":1,"title":["A comparison of various imputation algorithms for missing data"],"prefix":"10.1371","volume":"20","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-8370-6882","authenticated-orcid":true,"given":"J\u00fcrgen","family":"Kampf","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Iryna","family":"Dykun","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Tienush","family":"Rassaf","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2336-7991","authenticated-orcid":true,"given":"Amir Abbas","family":"Mahabadi","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"340","published-online":{"date-parts":[[2025,5,12]]},"reference":[{"key":"pone.0319784.ref001","volume-title":"Statistical analysis with missing data","author":"R Little","year":"2020"},{"key":"pone.0319784.ref002","doi-asserted-by":"crossref","DOI":"10.1201\/9781439821862","volume-title":"Analysis of incomplete multivariate data","author":"J Schafer","year":"1997"},{"key":"pone.0319784.ref003","doi-asserted-by":"crossref","DOI":"10.1002\/9780470510445","volume-title":"Missing data in clinical studies","author":"G Molenberghs","year":"2007"},{"key":"pone.0319784.ref004","volume-title":"Applied missing data analysis","author":"C Enders","year":"2010"},{"key":"pone.0319784.ref005","first-page":"375","article-title":"Are deep learning models superior for missing data imputation in surveys? Evidence from an empirical comparison","volume":"48","author":"Z Wang","year":"2022","journal-title":"Survey Methodol"},{"key":"pone.0319784.ref006","doi-asserted-by":"crossref","DOI":"10.1201\/9780429492259","volume-title":"Flexible imputation of missing data","author":"S Van Buuren","year":"2018"},{"key":"pone.0319784.ref007","first-page":"1","article-title":"mice: Multivariate imputation by chained equations in R","volume":"45","author":"S Van Buuren","year":"2011","journal-title":"J Stat Softw"},{"key":"pone.0319784.ref008","doi-asserted-by":"crossref","first-page":"287","DOI":"10.1080\/07350015.1988.10509663","article-title":"Missing-data adjustments in large surveys","volume":"6","author":"R Little","year":"1988","journal-title":"J Business Econ Stat"},{"key":"pone.0319784.ref009","unstructured":"Gaffert P, Meinfelder F, Bosch V. Towards an MI-proper predictive mean matching. 2016:15. https:\/\/citeseerx.ist.psu.edu\/document?repid=rep1&type=pdf&doi=5f13dd8d367c9b0ebc127c24487aa06773a2f00f."},{"key":"pone.0319784.ref010","doi-asserted-by":"crossref","first-page":"1070","DOI":"10.1093\/aje\/kwq260","article-title":"Multiple imputation for missing data via sequential regression trees","volume":"172","author":"L Burgette","year":"2010","journal-title":"Am J Epidemiol"},{"key":"pone.0319784.ref011","doi-asserted-by":"crossref","first-page":"92","DOI":"10.1016\/j.csda.2013.10.025","article-title":"Recursive partitioning for missing data imputation in the presence of interaction effects","volume":"72","author":"L Doove","year":"2014","journal-title":"Comput Stat Data Anal"},{"key":"pone.0319784.ref012","doi-asserted-by":"crossref","DOI":"10.1002\/9780470316696","volume-title":"Multiple Imputation for Nonresponse in Surveys","author":"D Rubin","year":"1987"},{"key":"pone.0319784.ref013","first-page":"13","article-title":"Multiple imputation for the fatal accident reporting system","volume":"40","author":"D Heitjan","year":"1991","journal-title":"J Royal Stat Soc Series C: Appl Stat"},{"key":"pone.0319784.ref014","doi-asserted-by":"crossref","first-page":"754","DOI":"10.1177\/0962280212465502","article-title":"A comparison of incomplete-data methods for categorical data","volume":"25","author":"D van der Palm","year":"2016","journal-title":"Stat Meth Med Res."},{"key":"pone.0319784.ref015","doi-asserted-by":"crossref","first-page":"162","DOI":"10.1080\/00031305.2016.1277158","article-title":"An empirical comparison of multiple imputation methods for categorical data","volume":"71","author":"O Akande","year":"2017","journal-title":"Am Statist"},{"key":"pone.0319784.ref016","doi-asserted-by":"crossref","first-page":"6310","DOI":"10.1002\/sim.3476","article-title":"The performance of multiple imputation for missing covariate data within the context of regression relative analysis","volume":"27","author":"R Giorgi","year":"2008","journal-title":"Stat Med"},{"key":"pone.0319784.ref017","doi-asserted-by":"crossref","first-page":"332","DOI":"10.1198\/016214504000001844","article-title":"Missing-data methods for generalized linear models \u2013 a comparative review","volume":"100","author":"J Ibrahim","year":"2005","journal-title":"J Am Stat Ass"},{"key":"pone.0319784.ref018","first-page":"1372","article-title":"Evaluation of multiple imputation with large proportions of missing data: how much is too much?","volume":"50","author":"J Lee","year":"2021","journal-title":"Iranian J Public Health"},{"key":"pone.0319784.ref019","volume-title":"A comparison of multiple imputation methods for missing covariate values in recurrent event data [Master Thesis]","author":"Z Huo","year":"2015"},{"key":"pone.0319784.ref020","doi-asserted-by":"crossref","first-page":"118","DOI":"10.1093\/ije\/dyp309","article-title":"Modelling relative survival in the presence of incomplete data: a tutorial","volume":"39","author":"U Nur","year":"2010","journal-title":"Int J Epidemiol"},{"key":"pone.0319784.ref021","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1177\/2050312118822912","article-title":"How handeling missing data may impact conclusions: a comparison of six different imputations methods for categorical questionnaire data","volume":"7","author":"M Stavseth","year":"2019","journal-title":"SAGE Open Med"},{"key":"pone.0319784.ref022","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1007\/s11135-021-01114-w","article-title":"How to choose an approach handeling missing categorical data: (un)expected findings from a simulated statistical experiment","volume":"56","author":"S Zhuchkova","year":"2022","journal-title":"Quality Quantity"},{"key":"pone.0319784.ref023","first-page":"7","article-title":"A comparison of strategies for missing data on machine learning classification algorithms","author":"T Makaba","year":"2019","journal-title":"2019 Int Multidis Inform Technol Engineer Conf"},{"key":"pone.0319784.ref024","first-page":"9","article-title":"Comparison of five imputation methods in handling missing data in a continuous frequency table","author":"M Mohammed","year":"2021","journal-title":"AIP Conf Proceed"},{"key":"pone.0319784.ref025","doi-asserted-by":"crossref","first-page":"764","DOI":"10.1093\/aje\/kwt312","article-title":"Comparison of random forest and parametric imputation models for imputing missing data using MICE: A CALIBER study.","volume":"179","author":"A Shah","year":"2014","journal-title":"Am J Epidemiol"},{"key":"pone.0319784.ref026","volume-title":"Classification and Regression Trees","author":"L Breiman","year":"1984"},{"key":"pone.0319784.ref027","volume-title":"The elements of statistical learning \u2013 data mining, inference, and prediction","author":"T Hastie","year":"2017"},{"key":"pone.0319784.ref028","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-030-56485-8","volume-title":"Random forests with R","author":"R Genuer","year":"2020"},{"key":"pone.0319784.ref029","doi-asserted-by":"crossref","first-page":"5","DOI":"10.1023\/A:1010933404324","article-title":"Random forests","volume":"45","author":"L Breiman","year":"2001","journal-title":"Machine learn"}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0319784","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,5,12]],"date-time":"2025-05-12T18:08:59Z","timestamp":1747073339000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0319784"}},"subtitle":[],"editor":[{"given":"Mar\u00eda Paula Fern\u00e1ndez","family":"Garc\u00eda","sequence":"first","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2025,5,12]]},"references-count":29,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2025,5,12]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0319784","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,5,12]]}}}