<![CDATA[The bank marketing department has collected data from bank customers or customers by marketing or socializing credit cards by telephone (telemarketing). Evaluation of credit card telemarketing that has been done by banks still lacks results and is efficient. One appropriate way to evaluate bank credit card telemarketing reports is to use data mining techniques. The purpose of using data mining is to determine the trends and patterns of customers who have the opportunity to subscribe to credit cards offered by banks. The research method uses Cross Industry Standard Process for Data Mining (CRISP-DM) with Genetic Algorithms for Feature Selection (GAFS) and Naive Bayes (NB). The results showed the number of attributes in the bank credit card telemarketing dataset were 15 attributes consisting of 14 ordinary attributes and 1 special attribute. The bank telemarketing dataset contains high-dimensional data, so the GAFS method is applied. After applying the GAFS method, it is obtained that 7 optimal attributes consist of 6 ordinary attributes and 1 special attribute. Six ordinary attributes include pekerjaan, balance, rumah, pinjaman, durasi, poutcome. While special attributes are target. The results showed the NB algorithm has an accuracy value of 86.71%. GAFS and NB algorithms increase the accuracy value to 90.27% for predictions of bank customers who take credit cards.]]>
Copyrights © 2019
