Enhanced early chronic kidney disease prediction using hybrid waterwheel plant algorithm for deep neural network...

Area of Science:

• Artificial Intelligence in Medicine
• Biomedical Data Science
• Computational Health

Background:

• Chronic Kidney Disease (CKD) affects millions globally, often diagnosed late due to limitations in traditional methods like Glomerular Filtration Rate (GFR) estimation.
• Early detection of CKD is crucial for effective management and preventing disease progression, but remains a significant clinical challenge.

Purpose of the Study:

• To develop and validate a novel deep neural network model for enhanced early prediction of Chronic Kidney Disease (CKD).
• To optimize the deep neural network using a hybrid metaheuristic algorithm combining Waterwheel Plant Algorithm (WWPA) and Grey Wolf Optimization (GWO).

Main Methods:

• The UCI CKD dataset was preprocessed using imputation, normalization, and synthetic oversampling to address data quality and class imbalance.
• A Multilayer Perceptron (MLP) regression model was trained and optimized using the proposed WWPA-GWO hybrid metaheuristic.
• Performance was benchmarked against other optimization algorithms (PSO, GA, WOA) and standard MLP.

Main Results:

• The WWPA-GWO optimized MLP model demonstrated significant improvements in prediction accuracy (e.g., R² = 0.9730) compared to the standard MLP (R² = 0.8793).
• The optimized model achieved reduced Mean Squared Error (MSE) and Root Mean Squared Error (RMSE), indicating higher precision.
• Computational time was substantially decreased with the optimized model (0.0999 s).

Conclusions:

• The WWPA-GWO hybrid optimization strategy is highly effective for enhancing deep neural network performance in early CKD detection.
• This approach offers a robust, efficient, and reliable tool for clinical application in identifying CKD at earlier stages.
• Future research directions include exploring advanced imputation, multi-modal data, and federated learning to further improve generalizability and clinical utility.