<![CDATA[The development of technology as an access to information about beauty products offered through internet is getting faster, especially about the review of beauty products that can help manufacturers to find out feedback about products from users, and help consumers to choose the appropriate beauty products easily. The product user can provide ratings and reviews on the sites that have been provided. Sometimes manufacturers and consumers have difficulty in differentiating and categorizing reviews into a rating as a determinant of the quality of a product. Therefore, a system is needed to simplify the right prediction of consumers or users of products on beauty products. In this study, a system was built using the calculation of the SO-CAL in an Inheritance-based method which applied on the K-NN algorithm and linear regression in the rating prediction. Results shows that the study using the SO-CAL in Inheritance-based method by testing using the Cross Validation / k-fold method obtained the average linear regression accuracy of 66% while the highest average accuracy of k-NN is 50% at Tolerance testing model 1. The average RMSE results in linear regression is 1.3628 while the k-NN algorithm is 2.1314. Hence, it can be concluded that the SO-CAL in Inheritance-based method is preferably applied to linear regression compared to the k-NN algorithm in the predicted rating.]]>
Copyrights © 2019
