Interpretable Feature Selection and Hybrid Deep Learning Models for Depressive Symptoms Prediction from Wearable Device Data.

Early detection and prediction of Depressive Symptoms is essential for improving mental health outcomes. This study proposes a hybrid deep learning and machine learning framework that utilizes tabular data collected from wearable devices, including sleep patterns, physical activity, and health-related indicators. Three ensemble learning models were used to identify influential predictors through the application of explainable artificial intelligence techniques such as SHAP and LIME. Based on the selected important features, two hybrid models were developed combining 1D-Convolutional Neural Network and Multi-layer Perceptron with LightGBM. The experimental results showed that models trained on selected features consistently outperformed those using the full feature set. The highest classification accuracy, 93.43%, was achieved by the multilayer perceptron model with LightGBM when trained on features selected by XGBoost. SHAP analysis highlighted the importance of features such as night sleep duration, age, and income responsibility, while LIME provided sample specific explanations that enhanced local interpretability. This framework enhances both the predictive performance and interpretability of depression prediction models and demonstrates the potential of wearable-derived behavior and physiological features as practical biomarkers for personalized risk assessment.