A Hybrid CNN–LSTM–GRU Deep Learning Model for the Accurate Classification of Chronic Kidney Disease

Abstract

Chronic Kidney Disease (CKD) is a progressive and often undiagnosed condition that poses a significant global health risk due to its silent progression and typical detection at a late stage. This study presents an advanced hybrid deep learning framework that integrates Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) architectures to improve the early prediction and classification of CKD. The framework employs a preprocessing pipeline that includes data cleaning, normalization, and class balancing using the Synthetic Minority Oversampling Technique (SMOTE) before performing deep feature extraction and sequence modeling. The hybrid CNN-LSTM-GRU model demonstrated outstanding performance, achieving an accuracy of 98.75%, a precision of 100%, a recall of 97.56%, an F1-score of 98.77%, and an Area Under the ROC Curve (AUC) of 0.988. These results significantly outperform conventional models such as LSTM, GRU, DNN, and 1D-CNN. The proposed framework has strong potential to support clinical decision-making systems for accurate, early CKD diagnosis, thereby improving patient outcomes and reducing the burden on healthcare systems.