Improved ResNet Models for Chronic Kidney Disease Prediction

Abstract

Chronic kidney disease (CKD), a consequential health issue that can deeply affect an individual's overall wellness, can be initiated by either kidney cancer or a gradual reduction in kidney function. As the chronic disease advances, it can reach a critical stage where only dialysis or surgery can save lives. Halting its progress becomes crucial. CKD patients also face a heightened risk of premature death. Early detection of associated conditions poses a challenging task for healthcare professionals aiming to prevent their onset. A unique deep learning model is presented in this work for the prediction of CKD. Many existing CKD prediction models have the drawbacks of producing less accuracy, mispredicting, utilizing more computation time, and using low-quality datasets or data with noise and missing values, leading to misprediction. So it is necessary to develop new techniques that give high predictions with less computation time. The objective of this research work is to build improved ResNet models for the prediction of chronic kidney disease and evaluate their performance in comparison to other cutting-edge machine learning and deep learning methods. This research work developed ResNet models such as improved ResNet 152v2 with inception, improved ResNet 101, and improved ResNet50 models that produced 99.90%, 96.53%, and 93.968% accuracy, respectively. The proposed ResNet models for CKD prediction will be useful to nephrologists and other medical professionals.