Estimating survival rate of kidney transplants by using data mining

todays, kidney failure is one of the costly problems of human society and use of renal replacement therapy is increasing in the world and Iran. Survival analysis is one of the fields in medical prognosis and data mining is a process of discovering unknown relationship and is a useful pattern from data and is known as a highly efficient method in survival analysis. Conclusively, the purpose of this study is predicting the survival of the kidney transplant patient's according to variables before kidney transplant.
In order to identify important factors for predicting survival in kidney transplant, informative requirements assessment was done by using self-designed questionnaire. Then, obtained information from the analysis of questionnaire was reviewed and data from 513 medical record of kidney patient in Sina Urology Research Center was extracted. Ultimately, by applying CRISP methodology, data mining was done by IBM SPSS Modeler 14.2 and C.5 algorithm.
In this study, BMI, ESRD and dialysis time were evaluated as the most effective factors in survival kidney transplant and extracted rules from the model can be used for predicting the survival of the transplanted kidney before the surgery. Accuracy rate of this model was estimated at 96.77%.
The high accuracy rate of C5.0 model shows the power of it in survival prediction. Furthermore, the most effective kidney transplant survival factors were identified and kidney transplanted survival of a new patient with distinctive features, can be predicted