PREDIKSI CUSTOMER CHURN PERUSAHAAN TELEKOMUNIKASI MENGGUNAKAN NAÏVE BAYES DAN K-NEAREST NEIGHBOR

Kaharudin Kaharudin (Universitas AMIKOM Yogyakarta)
Musthofa Galih Pradana (Universitas AMIKOM Yogyakarta)
Kusrini Kusrini (Universitas AMIKOM Yogyakarta)

Informasi Interaktif
Vol 4, No 3 (2019): Jurnal Informasi Interaktif

For a company it is very vital. Customers are the key to running a business that is run. But in reality there are loyal customers and some are churned out. Churn is defined as the tendency of customers to stop doing business with a company. It is important for companies to be able to identify customers who have a tendency to be churn customers. Then a report is needed to be able to identify and make decisions for management. Prediction method using Naïve Bayes method produces an accuracy of 76% and K-Nearest Neighbor produces information with a K = 1 value of 73%, K = 3 which is 76% and K = 5 by 78% It can be concluded that the K-Nearest Neighbor Method with K = 3 has a better value. The results of customer predictions for a company can be used to take an example for the customer so they will not churn.  Keywords: Prediction, Customer, Churn, Naïve Bayes, Telecomunication, K-Nearest Neighbor.