Hybrid Matrix-Ensemble Framework for Chronic Kidney Disease Diagnosis

Abstract

This study introduces a novel Hybrid Machine Learning and Mathematical Matrix Framework (HML-MMF) for optimizing the early detection of Chronic Kidney Disease (CKD), addressing critical limitations of conventional machine learning approaches, namely, poor interpretability, overfitting, and instability in clinical settings. The proposed framework uniquely integrates matrix algebra techniques, including column-wise mean imputation, Principal Component Analysis (PCA), Fisher Discriminant Analysis (FDA), and Min-Max normalization, with an ensemble of three classifiers: Gradient Boosting Machine (GBM), Support Vector Machine (SVM), and K-Nearest Neighbors (KNN). Applied to the UCI CKD dataset (400+ patients, 24 clinical features), the HML-MMF pipeline transforms raw clinical data into a mathematically structured, low-dimensional representation that enhances class separability while preserving physiological meaning. The final diagnosis is derived through soft voting, ensuring robustness and generalizability. Experimental results demonstrate that the hybrid model achieves 96.2% accuracy, 96.0% precision, 95.5% recall, 95.7% F1-score, and an AUC-ROC of 0.973, significantly outperforming both GBM-only (AUC = 0.942) and Matrix + SVM (AUC = 0.955) baselines. The scientific novelty lies in the synergistic fusion of interpretable matrix operations with ensemble learning, not as sequential steps, but as a unified architecture where mathematical transformations actively guide model optimization. This approach not only boosts performance but also provides clinical transparency, enabling practitioners to trace predictions back to key biomarkers like serum creatinine and BUN, which eigenvalue analysis confirms as dominant contributors. In the medical domain, this work offers a reliable, explainable decision-support tool that minimizes false negatives a critical requirement in CKD screening. Via bridging rigorous linear algebra with modern ML, the HML-MMF sets a new standard for trustworthy, high- erformance diagnostic systems in resource-constrained or data-imbalanced clinical environments.