SannyBold/malawi-mics-mental-health-ml

Machine learning analysis of childhood/adolescent depressive feelings using UNICEF Malawi MICS6 survey data, comparing logistic regression and random forest models in Python. # Predicting Depressive Feelings from UNICEF Malawi MICS6 Data This project applies supervised machine learning to UNICEF Malawi MICS6 survey data to investigate prediction of reported depressive feelings in children and adolescents using child, maternal and household variables. The analysis includes data preprocessing, handling of mixed categorical and numerical variables, exploratory analysis, and comparison of ridge logistic regression and random forest classifiers. Hyperparameters were tuned using stratified cross-validation and out-of-bag evaluation, with final performance assessed on a held-out test set. The random forest achieved a ROC-AUC of 0.65 and recall of 0.82, outperforming the logistic regression baseline across the main evaluation metrics. ## Repository contents - `malawi_mics_depression_classification.ipynb` – full analysis - `requirements.txt` – Python dependencies ## Data The underlying survey data are not included in this repository due to data access/privacy restrictions. ## Installation pip install -r requirements.txt