Data mining for chronic kidney disease prediction

Chronic Kidney Disease (CKD) is one of the most widespread illnesses in the United States. Recent statistics show that twenty-six million adults in the United States have CKD and million others are at increased risk. Clinical diagnosis of CKD is based on blood and urine tests as well as removing a sample of kidney tissue for testing. Early diagnosis and detection of kidney disease is important to help stop the progression to kidney failure. Data mining and analytics techniques can be used for predicting CKD by utilizing historical patient's data and diagnosis records. In this research, predictive analytics techniques such as Decision Trees, Logistic Regression, Naive Bayes, and Artificial Neural Networks are used for predicting CKD. Preprocessing of the data is performed to impute any missing data and identify the variables that should be considered in the prediction models. The different predictive analytics models are assessed and compared based on accuracy of prediction. The study provides a decision support tool that can help in the diagnosis of CKD.