<![CDATA[People are interested in buying and selling using e-commerce. In the place of buying and selling many products, one of which is beauty products. Many beauty products offer various advantages, with many products not being separated from reviews that assess a particular product regarding quality, advantages, disadvantages, and others. A review from a customer who has used the product can be used as a recommendation to choose the best product and as a determinant of product quality. Determination of product quality can be seen from various comments or reviews from customers to see whether the product is a best seller product or a less desirable product. Therefore, beauty product review data classification is carried out by labelling positive, negative, and neutral reviews. In this test using the Naive Bayes method with TF and TF Log weighting, TF weighting has a better accuracy value than TF Log weighting. TF weighting has an accuracy of 55% while TF Log weighting only has an accuracy of 52%. Meanwhile, to label the review using a kappa measure, in this test the kappa measure on each rater has the same value, namely 0.8.]]>
Copyrights © 2021
