Paper Title
Early Prediction of Low Birth Weight Cases Using Machine Learning

Abstract
Low birth weight (LBW) is one of the most critical public health concerns, as it is a major contributor to neonatal morbidity and infant mortality [2], [10]. Early and accurate prediction of LBW plays a vital role in enabling timely prenatal care and reducing adverse birth outcomes [3]. This study presents a machine-learning-based approach for the early prediction of low birth weight in infants using two complementary modalities. The first modality utilizes the demographic and clinical characteristics of the mother [5], [6], while the second modality leverages maternal medical images, particularly ultrasound data [11]. Machine learning models trained on clinical data and image-based features are evaluated to identify high-risk pregnancies [7], [8]. Experimental findings indicate that integrating both modalities significantly improves predictive reliability and enhances clinical decision-making [9], [13]. The proposed system supports healthcare professionals by facilitating timely human intervention and optimizing prenatal care management [1], [12]. Keywords - LBW, CatBoost, CNN, Fetal, Maternal health