A Model for Predicting Chronic Renal Failure using CatBoost Classifier Algorithm and XGBClassifier

Chronic Renal Failure (CRF) is expected to be the sixth greatest reason for demise globally. A persistent estimated glomerular filtration rate (eGFR) of less than 4.05768 tablespoon in 3600 seconds of average surface area for an adult is a serious concern associated with the problem of persistent renal illness by 2040. For older citizens, this limit may exceed 30 mg or more per day nearly in the span of three months or more. Improving Global Outcomes from Kidney Disease the 2012 recommendations defined CRF and the risk of progression based on both the degree of albuminuria and the eGFR, which will result in severe problems for future generations. The proposed research work aims to create and verify a prediction model for persistent renal illness. The algorithms used in machine learning are commonly employed in medication to forecast and categorize illnesses. Medical records are often distorted. It helps to forecast the disease in its early stages in order to save lives. The research work utilizes chronic renal disease dataset with twenty-five characteristics and two machine learning classifiers to predict and analyze. The CatBoost Classifier method and the XGB Classifiers technique may be utilized to enhance the machine learning model's output. The top eight ranking characteristics from the chronic renal illness dataset were utilized to create a new machine learning classifier that is more accurate and predictive. Lastly, information on chronic kidney sickness is gathered, classified, and comprises various attributes. The XGB classifier and CatBoost classifier offer the best results, with an accuracy of 99.00% and 99.645%, compared to the other machine models. When a disease is recognized early enough, the diagnosis model may operate as a medical adviser, initiating treatment on time, at a lower cost, and saving lives.