Leveraging KNN and XGBoost for Comprehensive Evaluation of Integration Effectiveness

Integration of diverse computational models has become a crucial approach to enhancing predictive accuracy and robustness. To evaluate the effectiveness of this integration, we propose a prediction method based on ensemble learning-the Enhanced Hybrid Stacking Model (EHSM). This method combines the strengths of the K-Nearest Neighbors (KNN) algorithm and the Extreme Gradient Boosting (XG-Boost) algorithm, using stacking techniques to enhance model performance. In this study, we conducted experiments using a dataset comprising various input features and target outcomes. The results show that the EHSM model excels in predicting target outcomes, achieving an AUC value of 0.95, which is significantly better than using KNN (AUC = 0.87) or XGBoost (AUC = 0.90) alone. By weightedly combining the predictions of KNN and XGBoost, the EHSM model effectively captures complex patterns in the data, reduces misclassification rates, and enhances prediction accuracy and robustness.