Predicting the Anxiety of Patients with Alzheimer’s Dementia using Boosting Algorithm and Data-Level Approach

Since overfitting due to imbalanced data can cause prediction errors during the learning process of machine and degrades performance model (e.g., sensitivity), it is necessary add an additional sampling technique in development step reduce overcome this issue, addition selecting a algorithm suitable for data. This study examined Alzheimer's patients living South Korea understand predictors anxiety using boosting algorithms (i.e., AdaBoost XGBoost) data-level approach (raw data, undersampling, oversampling, SMOTE) confirmed with best performance. We analyzed 253 elderly people who were diagnosed disease (aged from 60 74 years old) visited rehabilitation hospitals early dementia screening. developed models predicting XGBoost. Moreover, compared accuracy, sensitivity, specificity) models. The results showed that XGBoost based on SMOTE (accuracy=0.84, sensitivity=0.85, specificity=0.81) was identified as Consequently, presented SMOTE-XGBoost may provide higher accuracy than SMOTE-Adaboost developing outcome variable such future.