Analysis and prediction of chronic kidney disease

Abstract

In Sri Lanka, chronic kidney disease has become a significant public health problem over the past two decades. Since there are few signs or symptoms in the early stages, it is difficult to identify whether people have the CKD disease, because. Due to this reason, they do not get treatments. If the disease is detected at an early stage, CKD can be cured. Sri Lanka currently lacks comprehensive and systematic surveillance procedures to identify and monitor all aspects of CKD in the general population. The disease can be identified in the early stages if there is a proper dataset to analyze. Based on the data a predictive model can be developed and this will help doctors diagnose if a patient has early-stage CKD. CKD can prevent if this detects early and provide necessary treatments. As part of my research; I have developed a computerized system to capture and track aspects of CKD in Sri Lanka, including a predictive model to detect CKD in its early stages. The predictive model was developed using different types of data mining classification algorithms. In the healthcare sector, data mining is mainly used for disease detection. Broad data mining techniques exist for predicting diseases, such as classification, clustering, association rules, summaries, and regression. Additionally, the tool was developed to perform several analyses based on the collected data.