Novel machine learning approach for predicting Chronic kidney diseases

Abstract

Chronic Kidney Disease (CKD) is a chronic renal problem that affects the human kidney and makes it not to function properly or causes complete renal failure. It results in dialysis or causes other related diseases and reduces the quality of living. The symptoms of this disease cannot be identified in the preliminary stage. Only very lesser people are aware of this disease and can predict the symptoms at the earlier stage. However, it leads to prolonged disruption of kidney functional and finally causes it to failure and reduces the functionality completely. This can be occurred due to prolonged diabetes and also related with other diseases like Cardio-Vascular Disease (CVD). Due to inadequate prediction approaches lack of awareness in the preliminary stage, there is a delay in treating the patients’ at the initial phase of disease. From the various literature studies, it is identified that CKD can be predicted and treated in the earlier stage using the soft-computational techniques. Earlier CKD predictor model needs to be improved with higher prediction accuracy and precision. Therefore, there is a need for a decision support system that assists the nephrologists during the time of emergency conditions. Therefore, in this research, an efficient Clinical Decision Support System (CDSS) is modeled based on patients data to identify the occurrence of CKD and Non-CKD with the expert decision support system. Here, Machine Learning algorithms are used for designing the CDSS to predict CKD in prior stage.