Identifying Appropriate Feature Selection Techniques for Renal Disease Classification Models

Chronic kidney disease (CKD) or renal illness is the slow decline in the effective function of the kidney over many years. Heart and blood vascular disorders are both risk factors for kidney disease. The occurrence of these issues may be gradual and protracted. The progression of chronic renal disease can frequently be halted by early detection and treatment. Machine learning techniques can assist professionals in health care by intervening early by identifying indications and symptoms at an early stage, enhancing patient management and treatment outcomes. The proposed model compares random forest feature selection and boruta feature selection techniques and trains machine learning models, namely K-Nearest Neighbour (KNN), Naive Bayesian, and Logistic Regression (LR), with the selected features. Based on the resulting accuracies of the models for both feature selection techniques upon inference, it was found that Random forest feature selection provides better results. The potential advantages of utilizing cutting-edge technologies to assist in the prompt diagnosis and management of chronic diseases are brought out by these prediction approaches based on characteristics. Consequently, the objective is to determine whether the person has CKD or not by employing an effective number of features