A Comparative Study of Improved Ensemble Learning Algorithms for Patient Severity Condition Classification

Abstract

The evolution of Electronic Health Records (EHR) has facilitated comprehensive patient record-keeping, enhancing healthcare delivery and decision-making processes. Even with these developments, there are still certain difficulties when employing ensemble machine-learning techniques to analyze EHR data. This study aims to model the classification of patient severity using EHR data. Addressing issues with dimensionality and imbalance in EHR data. and to avoid overfitting by optimizing the ensemble model. The principal component analysis (PCA) method is used to address data dimensionality issues, and the synthetic minority oversampling technique (Smote) method is used to address data imbalance issues. After that, the ensemble model's hyperparameters are optimized using the Grid Search and Random Search approaches to prevent overfitting. In light of the study findings, the ensemble model's accuracy significantly improves after correcting data imbalance and dimensionality reduction. Notably, the Gradient Boosting Machine (GBM) and CatBoost models exhibited superior performance with an accuracy of 73%, achieved through experiments involving dimensionality reduction and handling of imbalanced data. Furthermore, optimization techniques such as Grid Search and Random Search were employed to enhance the EML models. The results of model optimization revealed that the GBM + Random Search model performed the best, achieving an accuracy of 74%, followed by the XGBoost + Grid Search model with an accuracy of 73%. The GBM model also excelled in distinguishing between positive and negative classes, boasting the highest Area under Curve (AUC) value of 0.78, indicative of its superior classification capabilities compared to other models. The study's findings offer a precise severity classification that medical professionals and teams can use to make quicker and more informed clinical decisions based on a patient's condition.