<![CDATA[Customer satisfaction is an important thing that is the company's goal. Whether you realize it or not, customer opinions written on social media, a little or a lot, will have an impact on potential customers. In its development, sentiment is increasingly easy to find in various online media, one of which is shopee. Product reviews are a very important source of information related to product quality and are very influential, both for consumers and producers. With a large amount of data for each product in the Shopee store, analyzing and concluding product review information will definitely take a lot of time if done manually. To overcome this, a sentiment analysis system is needed that can automatically extract important information that can objectively determine product quality and handle large textual information. The sentiment analysis system consists of several stages, namely crawling, pre-processing, word weighting, and sentiment classification. By applying the Naïve Bayes algorithm through the selection of range and frequency features, accuracy, accuracy and recall results will be obtained using the Confusion Matrix test. The API is used to collect data and perform simulations using data reviews on the Shopee website to generate positive and negative sentiments. The test will display the results of the percentile values for accuracy and retrieval. From the test results obtained results of accuracy of 99.5%, precision of 99.49%, recall of 100%. Thus it can be concluded that the classification method of the Naïve Bayes algorithm is quite relevant even though the accuracy is not yet 100%. The results of this study can provide solutions for Shopee stores to improve the quality of their sales.]]>
Copyrights © 2023
