Ensemble Model for the Risk of Anemia in Pediatric Patients With Sickle Cell Disorder

Abstract

Anemia is a major cause of morbidity and mortality of SCD patients in many parts of the world with the burden much higher in Sub Saharan Africa. This study developed an ensemble of machine learning algorithm for the prediction of the risk of anemia in pediatric SCD patients. Data for this study was collected from 115 pediatric SCD outpatients receiving treatment at a tertiary hospital in South-Western Nigeria. This study adopted a stack-ensemble model composed of deep neural network (DNN), multi-layer perceptron (MLP), and support vector machines (SVM) as base and meta-classifiers using the WEKA software. The ensemble models were compared following the stack-ensemble developed using SVM as a meta-classifier had the best performance with an accuracy of 72.7%. The study concluded that information about socio-demographic and clinical data can be used to assess the risk of anemia among SCD patients.