Related Experiment Video
Updated: Sep 4, 2025
Quantified Assessment of Infant's Gross Motor Abilities Using a Multisensor Wearable
Published on: May 17, 2024
Infant birth weight estimation and low birth weight classification in United Arab Emirates using machine learning
Wasif Khan1,2, Nazar Zaki1,2, Mohammad M Masud2,3
1Department of Computer Science and Software Engineering, College of Information Technology, United Arab Emirates University, 15551, Al Ain, United Arab Emirates.
Insights
Accurate prediction of infant birth weight (BW) is vital for newborn health. Machine learning models, particularly Random Forest for estimation and Logistic Regression with SMOTE for classification, show high performance in predicting low birth weight (LBW).
Area of Science:
- Medical Informatics
- Machine Learning in Healthcare
- Neonatal Health
Background:
- Accurate prediction of newborn birth weight (BW) is critical for assessing infant health and identifying risks associated with low birth weight (LBW).
- Machine learning (ML) has emerged as a powerful tool in medical diagnostics, with various systems proposed for LBW prediction using maternal features.
- Existing systems often differ in the maternal features utilized, necessitating a standardized approach for robust LBW classification and BW estimation.
Purpose of the Study:
- To establish a detailed framework for both birth weight (BW) estimation and low birth weight (LBW) classification using machine learning.
- To evaluate the performance of multiple ML algorithms on a comprehensive dataset incorporating various feature subsets and selection techniques.
- To investigate the impact of oversampling techniques, specifically the synthetic minority oversampling technique (SMOTE), on LBW classification accuracy.
Main Methods:
- A dataset comprising 88 features from 821 women across three hospitals in the United Arab Emirates was utilized.
- Thirty distinct machine learning algorithms were evaluated for both BW estimation and LBW classification tasks.
- Feature selection techniques and the synthetic minority oversampling technique (SMOTE) were employed, alongside five-fold cross-validation, to optimize model performance.
Main Results:
- The Random Forest algorithm achieved the best BW estimation, with a mean absolute error of 294.53 g.
- Logistic Regression, combined with SMOTE oversampling, yielded the highest classification performance, achieving 90.24% accuracy, 87.6% precision, 90.2% recall, and a 0.89 F1 score.
- Key maternal features such as diabetes, hypertension, and gestational age were identified as significant predictors for LBW classification.
Conclusions:
- Machine learning models, particularly Random Forest for estimation and Logistic Regression with SMOTE for classification, offer effective solutions for predicting infant birth weight and identifying low birth weight (LBW) risks.
- The study highlights the importance of specific maternal health indicators in the prediction of LBW.
- The findings provide a robust framework for developing advanced diagnostic tools to improve neonatal care and outcomes.
Abstract:
Accurate prediction of a newborn's birth weight (BW) is a crucial determinant to evaluate the newborn's health and safety. Infants with low BW (LBW) are at a higher risk of serious short- and long-term health outcomes. Over the past decade, machine learning (ML) techniques have shown a successful breakthrough in the field of medical diagnostics. Various automated systems have been proposed that use maternal features for LBW prediction. However, each proposed system uses different maternal features for LBW classification and estimation. Therefore, this paper provides a detailed setup for BW estimation and LBW classification. Multiple subsets of features were combined to perform predictions with and without feature selection techniques. Furthermore, the synthetic minority oversampling technique was employed to oversample the minority class. The performance of 30 ML algorithms was evaluated for both infant BW estimation and LBW classification. Experiments were performed on a self-created dataset with 88 features. The dataset was obtained from 821 women from three hospitals in the United Arab Emirates. Different performance metrics, such as mean absolute error and mean absolute percent error, were used for BW estimation. Accuracy, precision, recall, F-scores, and confusion matrices were used for LBW classification. Extensive experiments performed using five-folds cross validation show that the best weight estimation was obtained using Random Forest algorithm with mean absolute error of 294.53 g while the best classification performance was obtained using Logistic Regression with SMOTE oversampling techniques that achieved accuracy, precision, recall and F1 score of 90.24%, 87.6%, 90.2% and 0.89, respectively. The results also suggest that features such as diabetes, hypertension, and gestational age, play a vital role in LBW classification.
Related Concept Videos
01:17z Scores and Area Under the Curve
01:52Regression Toward the Mean