Comparison of Decision Tree and K-Means Clustering Algorithms to Determine Awards for Customer Loyalty

Abstract

PT Tangguh Buana Roda Indonesia has difficulty in retaining loyal customers due to less than optimal customer management. This research proposes the use of a data mining-based system to categorize loyal customers using the K-Means and Decision Tree methods. The evaluation shows that the combination of K-Means and Decision Tree algorithms provides a higher average accuracy of 93.7175%. Compared to using Decision Tree alone which reached 92.8525% and K-Means which was only 91.667%. With the combination of these two algorithms, it is expected to support the awarding of loyal customers and strengthen the relationship between customers and companies. The system that has been created is web-based which will facilitate strategic planning to increase customer loyalty.