APPLICATION OF NAÏVE BAYES CLASSIFIER ALGORITHM IN DETERMINING THE LEVEL OF CUSTOMER SATISFACTION WITH RUMBAI POST OFFICE SERVICES

Abstract

Technological advances make service and delivery of goods grow rapidly. Coupled with people's changing shopping habits by shopping online, shipping companies are increasingly needed. PT. POS Indonesia is the first expedition company in Indonesia. Currently PT. POS Indonesia has opened many POS Office branches in every region in Indonesia, one of which is the Rumbai POS Office located in Pekanbaru City. To continue to maintain the company while competing with other expeditions, the Rumbai POS Office must continue to maintain its customers by improving the quality of service. Survey analysis can be done to determine the extent of customer satisfaction with the services provided. To find out the level of customer satisfaction, you can use the classification method. Naive Bayes is a popular and effective machine learning algorithm for classification problems. The study used datasets sourced from the results of questionnaire distribution to customers of the Rumbai POS Office. The questionnaire used 14 indicators derived from the Community Satisfaction Index set by the Ministry of Agriculturein 2004. The classification resulted in a Satisfied class of 16 data with a percentage of 84.2% and a Dissatisfied class of 3 data with a percentage of 15.8%, it can be concluded that the service at the Rumbai POS Office is good. From the classification results, it is proven that the Naïve Bayes algorithm is able to predict well the level of customer satisfaction with an accuracy value of 94.74%, precision of 100%, and recall of 94.12%. The results of this research can later be used as information for the Rumbai POS Office to be able to improve service quality.