<p>Assessing the Association Between Maternal Age and Postpartum Depression Using Statistical Analysis and Explainable Artificial Intelligence</p>

Postpartum depression (PPD) can feel like a curse for a new mother in an otherwise happy family after childbirth. PPD can occur due to a wide range of factors, including physical and psychosocial issues. This research focuses on middle-aged mothers (40-45 years) to examine the prevalence of PPD. It further investigates whether age plays a crucial role in the occurrence of PPD. Although the initial dataset was unclassified, a popular unsupervised algorithm, K-means, was used to cluster the data. The result of the cluster for mothers aged 40-45 was further cross-checked using several statistical tests, including the chi-square test, the Mann-Whitney U-test, and logistic regression with odds ratios. While the Mann-Whitney U-test showed no significant difference in the severity of PPD between younger and older mothers, the chi-square test and logistic regression indicated that the age group 40-45 is a moderately crucial period for motherhood due to the risk of PPD. Finally, explainable AI was applied using a supervised algorithm (random forest) to assess the influence of age and other symptoms. This study clearly shows which symptoms are mainly responsible for PPD and which have a moderate effect. An accuracy of approximately 94% was achieved, with key symptoms including feelings of guilt, appetite problems, and irritability. The model achieved precision, recall, and F1-scores of 91.77%, 95.30%, and 93.50%, respectively. The receiver operating characteristic score was 98.40%.