Show simple item record

dc.contributor.authorKaratsiolis, Savvasen
dc.contributor.authorSchizas, Christos N.en
dc.contributor.editorKyriacou, Efthyvoulos C.en
dc.contributor.editorChristofides, Steliosen
dc.contributor.editorPattichis, Constantinos S.en
dc.creatorKaratsiolis, Savvasen
dc.creatorSchizas, Christos N.en
dc.description.abstractAn original classification algorithm is proposed for dealing with extremely imbalanced datasets that often appear in biomedical problems. Its originality comes from the way a neural network is trained in order to get a decent hypothesis out of a dataset that comprises of a huge sized majority class and a tiny size minority class. This situation is especially probable when forming machine learning databases describing rare medical conditions. The algorithm is tested on a large dataset in order to predict the risk of preeclampsia in pregnant women. Conventional machine learning algorithms tend to provide poor hypothesis for extremely imbalanced datasets by favoring the majority class. The proposed algorithm is not trained on the basis of the mean squared error objective function and thus avoids the overwhelming effect of the highly asymmetric class sizes. The methodology provides preeclampsia detection rate of 49% and normal case detection rate slightly above 76%. © Springer International Publishing Switzerland 2016.en
dc.publisherSpringer Verlagen
dc.sourceIFMBE Proceedingsen
dc.source14th Mediterranean Conference on Medical and Biological Engineering and Computing, MEDICON 2016en
dc.subjectLearning algorithmsen
dc.subjectArtificial intelligenceen
dc.subjectNeural networksen
dc.subjectClassification (of information)en
dc.subjectBiochemical engineeringen
dc.subjectMedical computingen
dc.subjectMean square erroren
dc.subjectLearning systemsen
dc.subjectObjective functionsen
dc.subjectMean squared erroren
dc.subjectClassification algorithmen
dc.subjectBiomedical problemsen
dc.subjectConventional machinesen
dc.subjectImbalanced Data-setsen
dc.subjectImbalanced datasetsen
dc.subjectMedical conditionsen
dc.subjectNeural networken
dc.titleVariable target values neural network for dealing with extremely imbalanced datasetsen
dc.description.endingpage522 Σχολή Θετικών και Εφαρμοσμένων Επιστημών / Faculty of Pure and Applied SciencesΤμήμα Πληροφορικής / Department of Computer Science
dc.type.uhtypeConference Objecten
dc.description.notesConference code: 172989</p>en
dc.contributor.orcidSchizas, Christos N. [0000-0001-6548-4980]
dc.contributor.orcidPattichis, Constantinos S. [0000-0003-1271-8151]
dc.contributor.orcidKyriacou, Efthyvoulos C. [0000-0002-4589-519X]

Files in this item


There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record