A new proposed feature selection method to predict kidney transplantation outcome

Kidney transplantation graft survival prediction is important because of the difficulty of finding the organs. The exact prediction of kidney transplantation outcome is still not accurate even with the enhancements in acute rejection results. Machine learning methods introduce many ways to solve the kidney transplantation prediction problem than that of other methods. The power of any prediction method relies on the choosing of the proper variables. Feature selection is one of the important preprocessing procedures. It is the method that selects the minimal suitable variables that introduced in a set of features. This paper introduced a new proposed feature selection method that combines statistical methods with classification procedures of data mining technology to predict the probability of graft survival after kidney transplantation. Univariate analysis using Kaplan-Meier survival analysis method combined with Naive Bayes classifier was used to specify the significant variables. Three data mining tools, namely naive Bayes, decision tree and K-nearest neighbor classifiers were utilized to examine the instances of kidney transplantation, and their accuracy was compared with using the new proposed feature selection method and without using it. Experimental results have presented that the new proposed feature selection method have better results than other techniques.