PERFORMANCE ANALYSIS OF C4.5 AND NAÏVE BAYES ALGORITHM ON CUSTOMER RELATIONSHIP MANAGEMENT (CRM) IN JATIMAS FURNITURE LTD.

Abstract

Customer relationship management (CRM) was a part of the marketing strategies to maintain customer satisfaction. Customer satisfaction was an important factor in a business to stay afloat or continued to grow so that business managers needed an evaluation from the customers to find out whether customers were satisfied or not satisfied with the services provided by the company under customer expectations. The objective of this study was to measure the customer satisfaction in the Jatimas Furniture Ltd. using C4.5 and Naïve Bayes Algorithm. The variables of this study were service, price, promotion, product quality, and facilities. This study was expected to assist the company in increasing the level of customer satisfaction to defend the company from its competitors. After that, the performance analysis of the C4.5 and Naïve Bayes Algorithm were done to find out the algorithm that had higher accuracy for evaluating customer satisfaction survey results. The tests were carried out through the Split Validation technique. The result of the tests 1 was that the test 1 used 6 variables of the Naïve Bayes Algorithm by generating 81.48% accuracy and AUC (Area Under Curve) was 0.871 through 90% training data and 10% testing data; while, the test 2 used 4 variables of decision trees C4.5 Algorithm by generating 76.19 % accuracy and AUC (Area Under Curve) was 0.500 through 70% training data and 30% testing data. The conclusion of this study was that the accuracy and AUC (Area Under Curve) of the Naïve Bayes Algorithm were higher than the decision tree C4.5 Algorithm.

Keywords: Customer Relationship Management (CRM), Decision tree C.45 Algorithm, Naïve Bayes Algorithm