The application of machine learning approaches to determine the predictors of anemia among under five children in Ethiopia

Abstract

Health professionals need a strong prediction system to reach appropriate disease diagnosis, particularly for under-five child with health problems like anemia. Diagnosis and treatment delay can potentially lead to devastating disease complications resulting in childhood mortality. However, the application of machine learning techniques using a large data set provides scientifically sounded information to solve such palpable critical health and health-related problems. Therefore, this study aimed to determine the predictors of anemia among under-5 year’s age children in Ethiopia using a machine learning approach. A cross-sectional study design was done using the Ethiopian Demographic and Health Survey 2016 data set. A two-stage stratified cluster sampling technique was employed to select the samples. The data analysis was conducted using Statistical Package for Social Sciences/SPSS version 25 and R-software. Data were derived from Ethiopian Demographic and Health Survey. Boruta algorism was applied to select the features and determine the predictors of anemia among under-5 years-old children in Ethiopia. The machine learning algorism showed that number of children, distance to health facilities, health insurance coverage, youngest child’s stool disposal, residence, mothers’ wealth index, type of cooking fuel, number of family members, mothers’ educational status and receiving rotavirus vaccine were the top ten important predictors for anemia among under-five children. Machine-learning algorithm was applied to determine the predictors of anemia among under- 5 year’s age children in Ethiopia. We have identified the determinant factors by conducting a feature importance analysis with the Boruta algorithm. The most significant predictors were number of children, distance to health facility, health insurance coverage, youngest child’s stool disposal, residence, mothers’ wealth index, and type of cooking fuel. Machine learning model plays a paramount role for policy and intervention strategies related to anemia prevention and control among under-five children.

Employing supervised machine learning algorithms for classification and prediction of anemia among youth girls in Ethiopia

“…Overall, the utilization of machine learning algorithms for classification and prediction offers numerous advantages, including automation, pattern recognition, adaptability, scalability, objectivity, handling non-linearity, feature selection, and generalization. These makes a powerful tool for addressing a wide range of real-world problems and driving data-driven decision-making 19 . Therefore, in this research, we have utilized eight advanced machine learning techniques, such as association rule mining, to forecast the condition of anemia by utilizing demographic health survey information.…”

Employing advanced supervised machine learning approaches for predicting micronutrient intake status among children aged 6–23 months in Ethiopia

“…Leveraging machine learning (ML) models can offer significant advantages and contribute to the existing empirical evidence and making the most accurate predictions enabling systems to learn from data rather than making prior assumptions ( 20 ). ML techniques excel in managing complex and nonlinear data, operate without preexisting assumptions, and capture intricate relationships among predictors ( 20 , 21 ). Besides, the previous study was confined with a limited ML algorithm, data balancing techniques, and small sample size.…”

Image_1.pdf

“…Leveraging machine learning (ML) models can offer significant advantages and contribute to the existing empirical evidence and making the most accurate predictions enabling systems to learn from data rather than making prior assumptions (20). ML techniques excel in managing complex and nonlinear data, operate without preexisting assumptions, and capture intricate relationships among predictors (20,21). Besides, the previous study was confined with a limited ML algorithm, data balancing techniques, and small sample size.…”