Segmentation and classification customer payment behavior at multimedia service provider company with K-Means and C4.5 algorithm
Segmentation and classification customer payment behavior at multimedia service provider company with K-Means and C4.5 algorithm
This research use model with join k-means segmentation and C4.5 classification algorithm because C4.5 weaknesses in difficulty to choose attributes. Be proven that extract customer potential attributes with k-means can help to increase C4.5 classification algorithm’s accuracy. This thing proved from the model accuracy increment from 59.02% to 77.31% and AUC from 0.537 to 0.836. Customer potential level can also be the reference in promotion, retention, and prevention of insolvency customer.