Identifying the Determinants of Depression among Ever Married Women in Bangladesh: A Machine Learning Approach

Abstract

Mental health remains one of the most challenging public health concerns in the world, particularly in a developing country like Bangladesh. The objective of this study is to determine the key determinants of depression among ever-married women in Bangladesh. This study used 19,987 ever-married women aged 15-49 years in Bangladesh, from the 2022 Bangladesh Demographic and Health Survey (BDHS). The key features selection was selected by using Boruta algorithm. Total seven machine learning approaches, including Extreme Gradient Boosting (XGBoost), Random Forest (RF), Decision Tree (DT), Support Vector Machine (SVM), Logistic Regression (LR), K-Nearest Neighbour’s (KNN), and Neural Network (NN) were evaluated for predictive performance. All models were highly tuned with 5-fold cross-validation and tested on several complementary predictions including accuracy, precision, recall, F1 score, Matthews Correlation Coefficient (MCC), Cohen's Kappa, and Area Under the Receiver Operating Characteristic Curve (AUROC). The top-performing algorithm was observed XGBoost approach with the highest AUROC of 0.799, indicating strong discriminatory power between non-depressed and depressed women.  Results from XGBoost approach revealed that anxiety, currently working, age group 45-49 years, internet user, menstruated last 6 weeks, pressure to pregnant, power of decision making etc. are most important determinants of the depression of married women in Bangladesh.

International Journal of Statistical Sciences, Vol. 25(2), November, 2025, pp 33-46