Risk Aware Birth Weight Prediction System

DOI

https://www.doi.org/10.59256/ijire.20260702032

Published In

Volume 7, Issue 2 (2026)

Published Online

March-April 2026

Pages

262-273

Article Type

Original Article

Check for updates

Correctly estimating fetal birth weight (FBW) before delivery is a clinically important problem, as it is directly associated with perinatal morbidity and long-term developmental outcomes.Low birth weight (LBW, < 2500 g) and macrosomia (HBW, ≥ 4000 g) are associated with increased neonatal mortality, complications during pregnancy, and a higher possibility of chronic diseases in later stages of life. Standard sonographic regression formulas demonstrate systematic errors for extreme weight categories due to class imbalance and intergroup morphological heterogeneity in clinical datasets [1], [15]. Current machine-learning (ML) methodologies primarily depend on cardiotocography (CTG) signals or specialized ultrasound biometrics (biparietal diameter, femur length, abdominal circumference), which are often inaccessible in primary-care antenatal environments [1]–[3]. This paper presents VOTEWEIGHT, a heterogeneous hardvoting ensemble that integrates predictions from Random Forest (RF) [16], Gradient Boosting (GB) [17], and Logistic Regression (LR) based on a specified set of twelve routine maternal clinical attributes consistently collected at each antenatal visit. The preprocessing pipeline involves six steps: a validation step, filling any missing values with the median, removing outliers based on the IQR, one-hot encoding, doing Min-Max Sinormalization and the train-test split. This verifies the accuracy of the data. Numerous tests confirm that VOTEWEIGHT has an overall accuracy of 95.4%, a macro-averaged F1 score of 0.952, a Mean Absolute Error of 193.7, and an RMSE of 256.1 g. It outperforms each and every individual base learner with statistical significance (McNemar’s test, p ¡ 0.01) [21]. Our exhaustive ablation study over the voting strategy, preprocessing options, and feature subsets shows that the primary predictive indicators are fasting glucose pre-gestational BMI and systolic blood pressure. VOTEWEIGHT functions as a real-time Flask web application, delivering clinician-facing predictions through a straightforward data-entry form. The system makes prenatal risk screening available to everyone by only using routine maternal data