{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,19]],"date-time":"2026-06-19T17:51:14Z","timestamp":1781891474067,"version":"3.54.5"},"reference-count":42,"publisher":"Oxford University Press (OUP)","issue":"11","license":[{"start":{"date-parts":[[2020,9,4]],"date-time":"2020-09-04T00:00:00Z","timestamp":1599177600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2020,11,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:sec>\n                  <jats:title>Objective<\/jats:title>\n                  <jats:p>Minority oversampling is a standard approach used for adjusting the ratio between the classes on imbalanced data. However, established methods often provide modest improvements in classification performance when applied to data with extremely imbalanced class distribution and to mixed-type data. This is usual for vital statistics data, in which the outcome incidence dictates the amount of positive observations. In this article, we developed a novel neural network-based oversampling method called actGAN (activation-specific generative adversarial network) that can derive useful synthetic observations in terms of increasing prediction performance in this context.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Materials and Methods<\/jats:title>\n                  <jats:p>From vital statistics data, the outcome of early stillbirth was chosen to be predicted based on demographics, pregnancy history, and infections. The data contained 363\u00a0560 live births and 139 early stillbirths, resulting in class imbalance of 99.96% and 0.04%. The hyperparameters of actGAN and a baseline method SMOTE-NC (Synthetic Minority Over-sampling Technique-Nominal Continuous) were tuned with Bayesian optimization, and both were compared against a cost-sensitive learning-only approach.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>While SMOTE-NC provided mixed results, actGAN was able to improve true positive rate at a clinically significant false positive rate and area under the curve from the receiver-operating characteristic curve consistently.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Discussion<\/jats:title>\n                  <jats:p>Including an activation-specific output layer to a generator network of actGAN enables the addition of information about the underlying data structure, which overperforms the nominal mechanism of SMOTE-NC.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Conclusions<\/jats:title>\n                  <jats:p>actGAN provides an improvement to the prediction performance for our learning task. Our developed method could be applied to other mixed-type data prediction tasks that are known to be afflicted by class imbalance and limited data availability.<\/jats:p>\n               <\/jats:sec>","DOI":"10.1093\/jamia\/ocaa127","type":"journal-article","created":{"date-parts":[[2020,6,3]],"date-time":"2020-06-03T11:12:26Z","timestamp":1591182746000},"page":"1667-1674","source":"Crossref","is-referenced-by-count":55,"title":["Synthetic minority oversampling of vital statistics data with generative adversarial networks"],"prefix":"10.1093","volume":"27","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-4116-344X","authenticated-orcid":false,"given":"Aki","family":"Koivu","sequence":"first","affiliation":[{"name":"Department of Future Technologies, University of Turku, Turku, Finland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mikko","family":"Sairanen","sequence":"additional","affiliation":[{"name":"PerkinElmer, Turku, Finland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Antti","family":"Airola","sequence":"additional","affiliation":[{"name":"Department of Future Technologies, University of Turku, Turku, Finland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tapio","family":"Pahikkala","sequence":"additional","affiliation":[{"name":"Department of Future Technologies, University of Turku, Turku, Finland"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2020,9,4]]},"reference":[{"issue":"5","key":"2020111714263946000_ocaa127-B1","doi-asserted-by":"crossref","first-page":"429","DOI":"10.3233\/IDA-2002-6504","article-title":"The class imbalance problem: a systematic study","volume":"6","author":"Japkowicz","year":"2002","journal-title":"Intell Data Anal"},{"key":"2020111714263946000_ocaa127-B2","doi-asserted-by":"crossref","first-page":"63","DOI":"10.1109\/TKDE.2006.17","article-title":"Training cost-sensitive neural networks with methods addressing the class imbalance problem","volume":"18","author":"Zhou","year":"2006","journal-title":"IEEE Trans Knowl Data Eng"},{"key":"2020111714263946000_ocaa127-B3","first-page":"73","author":"Ling","year":"1998"},{"key":"2020111714263946000_ocaa127-B4","first-page":"231","volume-title":"Encyclopedia of Machine Learning","author":"Ling","year":"2010"},{"key":"2020111714263946000_ocaa127-B5","volume-title":"Pattern Recognition and Machine Learning (Information Science and Statistics)","author":"Bishop","year":"2006"},{"key":"2020111714263946000_ocaa127-B6","author":"Weiss","year":"2007"},{"key":"2020111714263946000_ocaa127-B7","author":"Hoag","year":"2008"},{"key":"2020111714263946000_ocaa127-B8","doi-asserted-by":"crossref","first-page":"321","DOI":"10.1613\/jair.953","article-title":"SMOTE: synthetic minority over-sampling technique","volume":"16","author":"Chawla","year":"2002","journal-title":"J Artif Intell Res"},{"key":"2020111714263946000_ocaa127-B9","author":"Poolsawad","year":"2014"},{"key":"2020111714263946000_ocaa127-B10","first-page":"2672","author":"Goodfellow","year":"2014"},{"key":"2020111714263946000_ocaa127-B11","doi-asserted-by":"crossref","first-page":"101552","DOI":"10.1016\/j.media.2019.101552","article-title":"Generative adversarial network in medical imaging: a review","volume":"58","author":"Yi","year":"2019","journal-title":"Med Image Anal"},{"key":"2020111714263946000_ocaa127-B12","author":"Xu","year":"2018"},{"key":"2020111714263946000_ocaa127-B13","year":"2019"},{"issue":"9774","key":"2020111714263946000_ocaa127-B14","doi-asserted-by":"crossref","first-page":"1331","DOI":"10.1016\/S0140-6736(10)62233-7","article-title":"Major risk factors for stillbirth in high-income countries: a systematic review and meta-analysis","volume":"377","author":"Flenady","year":"2011","journal-title":"Lancet"},{"key":"2020111714263946000_ocaa127-B15","doi-asserted-by":"crossref","first-page":"f108","DOI":"10.1136\/bmj.f108","article-title":"Maternal and fetal risk factors for stillbirth: population based study","volume":"346","author":"Gardosi","year":"2013","journal-title":"BMJ"},{"issue":"11","key":"2020111714263946000_ocaa127-B16","doi-asserted-by":"crossref","first-page":"1177","DOI":"10.1093\/oxfordjournals.aje.a116620","article-title":"Risk factors for antepartum and intrapartum stillbirth","volume":"137","author":"Little","year":"1993","journal-title":"Am J Epidemiol"},{"issue":"3","key":"2020111714263946000_ocaa127-B17","doi-asserted-by":"crossref","first-page":"183","DOI":"10.1080\/14767050802559129","article-title":"Stillbirth in developing countries: a review of causes, risk factors and prevention strategies","volume":"22","author":"McClure","year":"2009","journal-title":"J Matern Fetal Neonatal Med"},{"key":"2020111714263946000_ocaa127-B18","doi-asserted-by":"crossref","first-page":"S5","DOI":"10.1186\/1471-2393-9-S1-S5","article-title":"Reducing stillbirths: screening and monitoring during pregnancy and labour","volume":"9 (Suppl 1","author":"Haws","year":"2009","journal-title":"BMC Pregnancy Childbirth"},{"issue":"1","key":"2020111714263946000_ocaa127-B19","doi-asserted-by":"crossref","first-page":"14","DOI":"10.1007\/s13755-020-00105-9","article-title":"Predicting risk of stillbirth and preterm pregnancies with machine learning","volume":"8","author":"Koivu","year":"2020","journal-title":"Health Inf Sci Syst"},{"issue":"5","key":"2020111714263946000_ocaa127-B20","doi-asserted-by":"crossref","first-page":"607","DOI":"10.1002\/uog.17290","article-title":"Prediction of stillbirth from maternal demographic and pregnancy characteristics","volume":"48","author":"Yerlikaya","year":"2016","journal-title":"Ultrasound Obstet Gynecol"},{"issue":"1","key":"2020111714263946000_ocaa127-B21","doi-asserted-by":"crossref","first-page":"274","DOI":"10.1186\/s12884-016-1061-2","article-title":"Predicting stillbirth in a low resource setting","volume":"16","author":"Kayode","year":"2016","journal-title":"BMC Pregnancy Childbirth"},{"issue":"3","key":"2020111714263946000_ocaa127-B22","doi-asserted-by":"crossref","first-page":"e0173461","DOI":"10.1371\/journal.pone.0173461","article-title":"A stillbirth calculator: development and internal validation of a clinical prediction model to quantify stillbirth risk","volume":"12","author":"Trudell","year":"2017","journal-title":"PLoS One"},{"issue":"1","key":"2020111714263946000_ocaa127-B23","doi-asserted-by":"crossref","first-page":"38","DOI":"10.1002\/pd.2644","article-title":"Prediction of miscarriage and stillbirth at 11-13 weeks and the contribution of chorionic villus sampling","volume":"31","author":"Akolekar","year":"2011","journal-title":"Prenat Diagn"},{"key":"2020111714263946000_ocaa127-B24","year":"2019"},{"key":"2020111714263946000_ocaa127-B25","volume-title":"Digital Design and Computer Architecture","author":"Harris","year":"2013","edition":"2nd ed."},{"key":"2020111714263946000_ocaa127-B26","first-page":"89","author":"Blagus","year":"2012"},{"key":"2020111714263946000_ocaa127-B27","first-page":"937","author":"Van Hulse","year":"2007"},{"key":"2020111714263946000_ocaa127-B28","author":"Arjovsky","year":"2017"},{"key":"2020111714263946000_ocaa127-B29","author":"Arjovsky","year":"2017"},{"key":"2020111714263946000_ocaa127-B30","author":"Gulrajani","year":"2017"},{"key":"2020111714263946000_ocaa127-B31","author":"Radford","year":"2015"},{"issue":"6789","key":"2020111714263946000_ocaa127-B32","doi-asserted-by":"crossref","first-page":"947","DOI":"10.1038\/35016072","article-title":"Digital selection and analogue amplification coexist in a cortex-inspired silicon circuit","volume":"405","author":"Hahnloser","year":"2000","journal-title":"Nature"},{"key":"2020111714263946000_ocaa127-B33","first-page":"972","author":"Klambauer","year":"2017"},{"key":"2020111714263946000_ocaa127-B34","author":"Maas","year":"2013"},{"key":"2020111714263946000_ocaa127-B35","first-page":"1026","author":"He","year":"2015"},{"key":"2020111714263946000_ocaa127-B36","volume-title":"Practical Methods of Optimization","author":"Fletcher","year":"1987"},{"key":"2020111714263946000_ocaa127-B37","author":"Kingma","year":"2014"},{"key":"2020111714263946000_ocaa127-B38","author":"Linnainmaa"},{"key":"2020111714263946000_ocaa127-B39","first-page":"2951","author":"Snoek","year":"2012"},{"key":"2020111714263946000_ocaa127-B40","first-page":"69","article-title":"Statistics corner: A guide to appropriate use of correlation coefficient in medical research","volume":"24","author":"Mukaka","year":"2012","journal-title":"Malawi Med J"},{"issue":"5","key":"2020111714263946000_ocaa127-B41","doi-asserted-by":"crossref","first-page":"569","DOI":"10.1001\/archpsyc.1983.01790050095011","article-title":"The predictive power of diagnostic tests and the effect of prevalence of illness","volume":"40","author":"Baldessarini","year":"1983","journal-title":"Arch Gen Psychiatry"},{"key":"2020111714263946000_ocaa127-B42","doi-asserted-by":"crossref","first-page":"613","DOI":"10.1002\/uog.17289","article-title":"Prediction of stillbirth from biochemical and biophysical markers at 11\u201313 weeks","volume":"48","author":"Mastrodima","year":"2016","journal-title":"Ultrasound Obstetr Gynecol"}],"container-title":["Journal of the American Medical Informatics Association"],"original-title":[],"language":"en","link":[{"URL":"http:\/\/academic.oup.com\/jamia\/article-pdf\/27\/11\/1667\/34363836\/ocaa127.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"http:\/\/academic.oup.com\/jamia\/article-pdf\/27\/11\/1667\/34363836\/ocaa127.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2020,11,17]],"date-time":"2020-11-17T19:27:22Z","timestamp":1605641242000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jamia\/article\/27\/11\/1667\/5901450"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,9,4]]},"references-count":42,"journal-issue":{"issue":"11","published-online":{"date-parts":[[2020,9,4]]},"published-print":{"date-parts":[[2020,11,1]]}},"URL":"https:\/\/doi.org\/10.1093\/jamia\/ocaa127","relation":{},"ISSN":["1527-974X"],"issn-type":[{"value":"1527-974X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2020,11]]},"published":{"date-parts":[[2020,9,4]]}}}