A Comprehensive Study on Postpartum Depression Prediction Using Machine Learning Approaches

Postpartum depression (PPD) presents a significant mental health concern for new mothers, often going undetected due to limitations in conventional screening methods like the Edinburgh Postnatal Depression Scale (EPDS). This project proposes a machine learning-based web application designed to automate PPD risk assessment. The system leverages a Feed-Forward Artificial Neural Network (FFANN) model trained on EPDS scores, achieving a prediction accuracy of 95%. Developed using Streamlit, the platform allows users to input their responses, visualize results via interactive charts, and download personalized reports in PDF format. A literature review of ten existing methods—ranging from traditional ML algorithms to deep learning and neuro-fuzzy models—was conducted for comparison. The system also includes mental health resources and a feedback mechanism, offering a comprehensive and accessible solution for early- stage PPD screening. The tool demonstrates the feasibility of integrating machine learning into maternal mental healthcare, aiming to improve timely intervention and support.