Predicting Chronic Kidney Disease using ML algorithms and XAIV

Abstract

Kidney disease is a significant health concern that is currently affecting individuals of all age groups. To predict the occurrence of chronic renal disease, a large number of scholars have employed machine learning and deep learning techniques. However, the efficacy of these methods is often hampered by a lack of transparency, which is a major issue in the application of artificial intelligence in healthcare and medical analysis. As such, the lack of clarity has prompted concern. To interpret the results of predictive models, the present study proposes the deployment of four machine learning algorithms, including Decision Tree, Logistic Regression, Multi-layer Perceptron Classifier, and Support Vector Machine, in combination with explainable AI (XAI) interface, leveraging the local interpretable model-agnostic explanation (LIME) and shapely additive explanation shapely additive values (SHAP). The proposed models are intended to facilitate effective decision-making in clinical research and therapeutic practices.