Hybrid Model Combining Deep Neural Network and XGBOOST for Predicting Chronic Kidney Disease
Background: chronic kidney disease is a global public health issues that affect approximately one billion people globally. kidney disease affect people in different ways across the world. Aim of this study is to enhance the prediction performance of chronic kidney disease (CKD) by developing a hybrid model that combing deep neural network and XGBoost algorithms, by utilizing a comprehensive set of features derived from clinical data, the model was designed to distinguish between healthy individuals and those at risk of CKD. The approach involved analyzing various metrics such as accuracy, precision, recall, and F1-score to evaluate model performance comprehensively. Model achieved an accuracy of 92%, with precision at 94% and recall at 92%, indicating its robustness in identifying positive cases while minimizing false positives. Data were collected from patient records, and a hybrid model was implemented, combining different algorithms to enhance predictive capabilities. The training process utilized multiple iterations to optimize the model's performance, resulting in a well-generalized tool for clinical practice. These findings suggest that the implementation of such a predictive model can significantly improve early detection and management of chronic kidney disease, thereby enhancing patient outcomes and resource allocation in healthcare settings. Future research should explore additional features and methodologies to further refine prediction accuracy and applicability in diverse clinical environments
Download Paper