An Improved Filter Based Feature Selection Model for Kidney Disease Prediction

In terms of societal health threats, kidney disease has been viewed as an increasing threat in the modern day. With an increasing incidence, chronic kidney disease (CKD) is a global public health concern. Early detection allows us to control the initial situation and administer treatment, but it also represents the most efficient means of addressing the growing global relevance sustainably. Accurately classifying kidney disorders plays a crucial role in clinical mining and is one of the most active study areas in medical data analysis. To anticipate an important and novel attribute that is not used in previous studies to diagnose, a new feature selection approach called WCR (Weighted Class Relief) was suggested in this work using the Relief feature selection model. The prediction accuracy of the proposed model is assessed using datasets related to kidney disease. Different metrics, including accuracy, precision and recall, are employed to evaluate WCR performance. The classification accuracy on MULTISurf, Relief-F, and Relief has been investigated using four classifiers, including SVM, KNN, Random Forest, and Naive Bayes. The outcome demonstrates that the suggested attribute selection strategy is successful and efficient at identifying the kidney disease.