<![CDATA[Twitter has a major role in the development of social, communication, psychological, marketing and political aspects. Posts Tweet comments or review indirectly will be a review of the assessment on a product. One of the most sought after products sectors today is beauty and skin care products. They look for products that they share with others, so they have a picture that affects their interest on the opinions of others who delivered via Twitter related results after using the product. Sentiment analysis can help in analyzing and classifying into positive and negative terms of twitter-related opinions about product trends and product quality in the public view. Opinions and comments related to Mustika Ratu's products are the subject of this study, citing the economic growth and the large number of users of Musitka Ratu who are companies in the field of beauty skin and beauty care. The Naive Bayes Classifier method is selected for implementation use, and has a fast performance in training, while the addition of Information is required for the feature selection process by reducing the presence of irrelevant words in the data used. The test is performed with 200 data (100 positive documents, and 100 negative documents) using the thresholds : 0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, dan 0.10. The results obtained are adjusted for a difference of 4%, the highest average value if no Information Gain (threshold 0) is 70%, while using Information Gain (threshold 0.01) equal to 74%. This is influenced by several factors such as the amount of data and data that spread from data data and documents. The highest accuracy value is obtained at K1 (threshold 0,02), then K5, K6 (threshold 0.01), and K7 (threshold 0,02 and 0,08) with percentage 85%, while at k with threshold at the lowest point 50%.]]>
