Publication

Abstract : Chronic Kidney Disease (CKD) characterizes a substantial problem on the healthcare system due to the growing number of patients, high risk of evolution to end-stage renal syndrome, and deprived prediction of illness and mortality. The goal of this project is to create a machine-learning model that leverages comorbidity and medication data from UCI Database to predict the development of CKD in the following 6 to 12 months, and hence the prevalence of the disease in the community. The proposed system predicts the CKD with minimal feature input instead of dumping all the features which may not relevant to predict the disease. To achieve this we have planned to approach by three feature selection algorithm with combination of two feature Extraction algorithm. After performing feature selection and Feature Extraction, those features will be trained with different Machine Learning algorithm. The accuracy of best combination algorithm will be implemented for predicting the CKD in a more efficient manner.