BamClassifier: a machine learning method for assessing iron deficiency
Résumé
Iron deficiency (ID) is a well-known cause of anaemia and could lead to adverse clinical and functional impairments. However, ID is under-diagnosed due to non-specific symptoms, difficulties in interpreting ambiguous assessment outcomes and suboptimal sensitivities of methods in some circumstances. In this study, we present BamClassifier, a machine learning method for assessment of ID. This method proceeds by repeated selection of samples of instances from routine complete blood count data in such a way that each observation is included in exactly one sample. Then, a median-supplement machine learning model built from each sample, and the performance of the model on test instances are aggregated into a bag of predictions from which ID statuses are assigned to samples by way of the highest frequency counts. We show the effectiveness of our method by applying to real datasets obtained from different investigations in Ghana and simulated data as well. Our method obtained perfect area under receiver operating characteristic curve in all experiments and significantly outperformed other well-established methods in terms of accuracy, sensitivity, specificity, precision, and diagnostic odds ratio in all our evaluations. A successful application of the method will permit the study of large collections of samples for ID assessments, save time and cost using complete blood count parameters, while standardizing interpretation of outcomes of such investigations.
Citer ce document
Exporter : BibTeX · RIS (Zotero, Mendeley, EndNote)
Accès au document
Texte intégral archivé sur SPHAERO, en accès ouvert : lisible gratuitement en ligne après connexion. Le document reste aussi disponible à sa source. Se connecter
Voir l'article sur le site de la revueLicence et provenance
Licence : CC BY
Notice moissonnée depuis OpenAlex le 29/09/2026. Le document reste hébergé par sa source.
Voir le document à la source →