Preterm Birth Prediction using Machine Learning

Preterm birth, a global maternal-fetal health challenge, is addressed through Electro-hysterography (EHG) for non-invasive uterine contraction monitoring. employs K- Nearest Neighbors (K-NN) and Logistic Regression for comparative analysis in preterm birth prediction using EHG data. Phase one includes acquiring and preprocessing EHG signals from pregnant individuals to capture vital temporal patterns. Feature extraction techniques derive relevant characteristics for predictive modeling. K-NN, emphasizing simplicity and efficiency, predicts preterm birth risks based on data point proximity. Simultaneously, Logistic Regression models the relationship between EHG features and preterm birth outcomes. Model performance is systematically evaluated using sensitivity, specificity, and accuracy metrics. Comparative analyses between K-NN and Logistic Regression inform on algorithm strengths and limitations. contributes to refining predictive models and guiding future algorithm selections for improved prenatal care and timely interventions, ultimately advancing understanding and mitigating preterm birth risks.