Improved CKD classification based on explainable artificial intelligence with extra trees and BBFS

Abstract

Chronic kidney disease is a persistent ailment marked by the gradual decline of kidney function. Its classification primarily relies on the estimated glomerular filtration rate and the existence of kidney damage. The kidney disease improving global outcomes organization has established a widely accepted system for categorizing chronic kidney disease. explainable artificial intelligence for classification involves creating machine learning models that not only accurately predict outcomes but also offer clear and interpretable explanations for their decisions. Traditional machine learning models often pose difficulties in comprehending the intricate processes behind specific classification choices due to their intricate and obscure nature. In this study, an explainable artificial intelligence-chronic kidney disease model is introduced for the process of classification. The model applies explainable artificial intelligence by utilizing extra trees and shapley additive explanations values. Also, binary breadth-first search algorithm is used to select the most important features for the proposed explainable artificial intelligence-chronic kidney disease model. This methodology is designed to derive valuable insights for enhancing decision-making strategies within the field of classifying chronic kidney diseases. The performance of the proposed model is compared with another machine learning models, namely, random forest, decision tree, bagging classifier, adaptive boosting, and k-nearest neighbor, and the performance of the models is evaluated using accuracy, sensitivity, specificity, F-score, and area under the ROC curve. The experimental results demonstrated that the proposed model achieved the best results with accuracy equals 99.9%.

Ähnliche Arbeiten