Screening of Chronic Kidney Disease Patients System Based on Data Mining Techniques

Abstract

The purposes of the research were to developed a Screening of Chronic Kidney Disease Patients System Based on Data Mining Techniques, compare the performance of algorithm of predicting disease and evaluate the performance of predicting disease system. The main functions of this system included the management of information of patients and medical personnel, physical examination, prediction of kidney disease, assessment of predictive result, model management and other various reports.
The research findings showed that the Random Forest Technique that utilized the dataset containing Oversampling Data had the highest accuracy value of all the information was 97.29%, precision value was 95.76% and F-Measure value was equal to 97.44% and this (particular) data mining technique can be used to develop a model to predict the occurrence of kidney disease. The result of performance evaluation for the prediction of kidney disease by the system found that (the system) is able to predict the new data as correct as correct as 95.71%. It can be concluded that the techniques and model can be applied in the development of medical support systems effectively in the future.