PREDIKSI TINGKAT LOYALITAS PELANGGAN MENGUNAKAN ALGORITMA C4.5 BERBASIS BACKWARD ELIMINATION

Syaifuddin Syaifuddin
Purwanto Purwanto
Catur Supriyanto

Jurnal Teknologi Informasi Cyberku
Vol 11 No 2 (2015): Jurnal Teknologi Informasi CyberKU Vol.11 no 2

Customer loyalty is one of the capital to maintain the company's business strategy in the long run. In thelast two decades of Customer Relationship Management (CRM) has grown to become one of the majortrends in marketing, both in education and in the world practice. CRM is a comprehensive businessstrategy of a company that enables the company to effectively manage the company's relationship with thecustomer. Automatic feature selection algorithm is used with the aim of selecting a subset of the featuresin the dataset in order to reach the maximum level of accuracy in classification. The use of data miningtechniques to predict customer loyalty combines C4.5 algorithm with feature selection BackwardElimination. C4.5 algorithm based backward elimination can improve the accuracy in the prediction ofcustomer loyalty, compared with C4.5 algorithm without feature selection. C4.5 algorithm basedbackward elimination generate income per month attribute, type of subscription, registration fee, the costof the bill, and the old subscription