Prediction of Parkinson’s Disease Depression Using LIME-Based Stacking Ensemble Model

Depression symptoms are comparable to Parkinson’s disease symptoms, including attention deficit, fatigue, and sleep disruption, as well of dementia such apathy. As a result, it is difficult for caregivers diagnose depression early. We examined LIME-based stacking ensemble model predict the patients with disease. This study used epidemiologic data (EPD) from Korea Disease Control Prevention Agency’s National Biobank, which included 526 patients’ information. Logistic Regression (LR) meta-model, five base models, LightGBM (LGBM), K-nearest Neighbors (KNN), Random Forest (RF), Extra Trees (ET), AdaBoost. After cleansing data, was trained using 261 participants’ 10 variables. According research, best combination ET + LGBM RF LR, harmonious model. In order achieve prediction explainability, we also combined explainable can help identify start treatment on them early in way that medical professionals comprehend.