OPINION MINING ABOUT PARFUM ON E-COMMERCE BUKALAPAK.COM USING THE NAÏVE BAYES ALGORITHM

Tjut Adek, Rizal and Fikry, Muhammad and Annisa, Helmina (2020) OPINION MINING ABOUT PARFUM ON E-COMMERCE BUKALAPAK.COM USING THE NAÏVE BAYES ALGORITHM. JITK (Jurnal Ilmu Pengetahuan dan Teknologi Komputer), 6 (1). pp. 107-113. ISSN 2685-8223, 2527-4864

Preview

Text Jurnal JITK Rizal.pdf - Published Version Download (1MB) | Preview

Abstract

Information plays a very important role in the rapid development ofthe world. Many people use online media to search for information, one of which is to find out information about the negative or positive of a product in e-commerce based on the comments that exist. To find out the classification of all comments-comet takes quite a long time in reading it. So,to make it easier than that all made a classification system to determine the classification of comments. In this classification process, the Naive Bayes algorithm is used as a solution to the problem. The process with the Naïve Bayes algorithm requires training data which is used as learning material from the system. The training data used is taken from one e-commerce site, Bukalapak.com regarding perfume products. Taking comments from Buakalapak.com used crawling techniques to retrieve comments from the whole product. The training data needed in this system is 1000 comments consisting of 500 positive training comments and 500 negative training comments. To get the accuracy value, it requires 300 test comments consisting of 150 positive test comments and 150 negative test comments. From the results of testing with Naive Bayes, the accuracy rate can be quite good, namely with a precision value of 96.44%, 96.34% recall,and an accuracy of 96.33%.

Actions (login required)

View Item