Data Mining Application in Prediction of Potential Customers of POS Machine Users in Fund Transaction

摘要

The main idea of the paper is predicting potential users of POS machine in bank system by Application of k-means and J48, classification algorithms. The card holders customer data along with book information have been collected, cleaned, integrated and transformed for testing using the clustering and classification algorithms. K-means clustering applied and used as an input for classification. Algorithm are tested certain cross validation and splitting the dataset into some percent for training and remaining for testing, techniques by setting the cluster file formed by the cluster models as dependent variables and the remaining variable as independent variables. Lastly, a comparison of the J 48 and Naïve Bayes Classification models in terms of the overall classification accuracy in classifying high level customers and accuracy in classifying low level/value customers have been undertaken. Therefore, clustering and Classification models have been the best in these evaluation parameters and thus one of algorithms selected as a better classifier in an electronic fund transaction in POS machine.