Detection of Chronic Kidney Disease using Machine Learning Algorithms

Abstract

The diagnosis of kidney disease, often called chronic renal illness, is known as chronic kidney disease (CKD). Chronic kidney disease (CKD) is a widespread and chronic health issue that requires preventative measures for early identification in order to slow the disease's progression. Using complex techniques like Support Vector Machine (SVM) and Logistic Regression, this study explores the field of machine learning. For robust model construction, use Random Forest and Decision Tree. Through the use of Variance Inflation Factor (VIF) for feature engineering, the dataset is carefully refined to improve the visibility of relevant features. To address any possible data imbalance concerns, Synthetic Minority Over-sampling Technique (SMOTE) is also utilized, promoting a fairer representation of classes in the dataset. By means of a thorough assessment procedure, the study methodically pinpoints the key elements that contribute to a precise diagnosis of chronic kidney disease. Each algorithm's effectiveness is evaluated using performance indicators like recall, accuracy, precision, and Fl-score.