An Intelligent Decision Support System for Early Detection of Chronic Kidney Disease using Machine Learning Models

All Chronic Kidney Disease (CKD) has become one of the rapidly increasing and frightening non – communicable disease in the world. $10 \%$ of the population of India and more than $10 \%$ of the population of the world are affected by CKD. The population mentioned here covers people in between 20 to 80 years of age. There are even children and adolescents in quite a big number affected by CKD but not included in this percentage. So many factors are responsible for this. If this disease is detected in advance, there are many treatments available to give the good quality of life back to the affected people. The work presented here is about predicting the CKD in advance based on the clinical parameters of the patients. The Machine Learning algorithms namely Logistic Regression, Randomizable Filtered Classifier and Reduced Error Pruning (REP) Tree are implemented over a dataset with 24 parameters of 400 patients and the prediction percentages are compared. Further the results of the best classification technique is improved by applying Ensemble Bagging Algorithm. It has given a result of $\mathrm{9 9 . 2 5 \%}$ accuracy in prediction. The procedure is carried out for the dataset with feature selection procedure done in initial stage.