Indonesia Conference Directory

Determining possible technology used in calculating customer buying interest is important for the company as it relates to marketing strategies and policies. The Naïve Bayes Classifier is a possible method to measure customer buying interest. This study aims to analyze the accuracy of this method in calculating customer buying interest of internet token. Data were obtained from past sales and then analyzed using the Naïve Bayes Classifier to predict future opportunities by defining the class of attributes. The classifications used were operator, internet quota, token active period, and product price. This study found that the Naïve Bayes Classifier was provably accurate in calculating customer buying interest as well as comparing the prediction with actual results aside the data training. Results showed that this method successfully classifies 10 products of 858 transactions as data training and 10 products of 115 transactions as data test. Here indicates that all data test and data training analysis showed eight of the tenth meaning that eight of ten predictions were correct or it had 80% accuracy. This study is expected becoming a reference in analyzing data sales and predicting future sales conditions, so it will help to determine appropriate strategy of product development.