Comparison of Classification C4.5 Algorithms and Naïve Bayes Classifier in Determining Merchant Acceptance on Sponsorship Program

Abstract

The large number of merchants that make sponsorship held by the Bank reaches thousands, data mining is used to classifying thousands of data. Naïve Bayes algorithm and C 4.5 are classification algorithms in data mining. The classification results are used as determinant where the merchant deserves to receive the sponsorship program, which potentially provides the source of funds and increase the brand awareness of the company by looking at the performance, transaction amount, total nominal, average daily transaction, average transaction nominal. Comparison results show that TheC 4.5 algorithm is the best model for handling case of Merchant eligibility in the Sponsorship Program. This can be proved by looking at the level of accuracy generated on the testing and validationprocess of the model. Both models have the same AUC value but the C 4.5 algorithm produces a superior accuracy value with a difference of 0.45% compared to Naïve Bayes.