Data mining classification algorithms in credit risk

Nowadays Big Data and its applications have become very important because of the size of the information it has contained. Especially, banking and insurance industries use big data analysis, data mining analysis, to determine the default risk. In this study, the data which was taken from Turkish Statistical Institution, containing 22745 instances and 14 attributes, was used. The purpose of this study is to select classification algorithm that can help to determine default risk of individuals according to their characteristics. Bayes Network, Naive Bayes, Logistic Regression, J48, Random Forest, Multi Layer Perceptron classification algorithms were selected and compared by using accuracy, root mean squared error, Roc curve , recall, and precision statistics and the best classification algorithm was determined. To do analysis WEKA an open source data mining program was used.