ARCHIVES

Postpartum depression (PPD) is a prevalent yet often undiagnosed mental health challenge. Postpartum women experience, which can negatively affect a woman's mental health, infant development and family health outcomes. PPD can continue to be a problem that is not always easily identified, especially in the presence of psychological and social/behavioural factors. This paper introduces a deep learning framework to predict postpartum depression from structured psychometric and demographic data from the Edinburgh Postnatal Depression Scale (EPDS). A dataset of 4,200 data records was used that included responses from the EPDS questionnaire and behavioral items including stress level, sleep duration and social support. To categorize PP depression risk, 5 deep learning architectures were created and compared. Experimental results show that the neural network-based models can successfully learn the complex relationships in mental health data and show acceptable prediction performance. The designed framework illustrates how AI can be effectively utilized in supporting early screening and decision support in maternal health care environments.