Analisis Data Bank Direct Marketing dengan Perbandingan Klasifikasi Data Mining Berbasis Optimize Selection (Evolutionary)

Abstract

In determining marketing strategies, the bank performs a classification from a customer database, the database will be analyzed by a decision maker and this is not easy for a decision maker, because of the complexity of the vast data and the many attributes of the data owned, so that it becomes an obstacle and obstacle. in decision making. This of course can have a negative effect on the company's business processes because there will be delays in determining marketing strategies. Data mining method is a method that can classify large data to determine the level of accuracy of a database. In overcoming these problems, it is necessary to do a database analysis to determine the accuracy level of the database classification owned by the company. For this reason, in this study a classification process will be carried out with the Bank Direct Marketing dataset taken from the UCI Machine Learning Repository web, using the Naïve Bayes algorithm, K-Nearest Neighbor, Support Vector Machine with Optimize Selection (Evolutionary) optimization, the calculation process using a data mining application. namely Rapidminer 5.3, to find the highest accuracy value from the calculation algorithm. Test method with 10-fold cross validation. In this study, the classification results with the highest level of accuracy were obtained using Optimize Selection (Evolutionary) optimization, namely the Naïve Bayes algorithm 90.18%, then K-Nearest Neighbor 86.66%, and Support Vector Machine 89.40%.