<![CDATA[Crowded producers of beauty product produce good and varied products. This has attracted consumers to use these beauty products. More and more consumers are using these beauty products, making producers try various innovations on their products. Innovation can be obtained from many comments, advices, or reviews made by consumers on variety of products. Benefits of product reviews for consumers are also useful to obtain information before buy a product. Many results of the review are not accompanied by rating. This makes it difficult for producers to classify reviews into certain sentiments. In this research aims to classify review into certain sentiments automatically into rating. In this research built a system using Semantic Orientation Calculator and Linear Regression methods. Breaking sentences in a review into n-gram (bigram and trigram) and one sentence aims to improve the results of predictions. Results of testing on this system are 23%, 71%, 67% on accuracy of bigram, 24%, 71%, 67% on accuracy of trigram, and lowest 24%, 67%, 64% on accuracy of one sentence with tolerance 0, tolerance 1, and sentiment reviews. The best result of testing on breaking sentence using n-gram (bigram and trigram) was good enough to solve problem in this research.]]>
Copyrights © 2019
