Sentiment Analysis Of Fiverr Application Reviews Using TF-IDF Feature

Most people are interested in being a freelancer. This happens because of the rapid development of technology, making it easier for people to move and providing many choices in determining the type of work. One of the most popular freelance apps is the Fiverr App. The Fiverr application has received many reviews from its users, both positive, negative, and neutral. This study aims to obtain the results of sentiment classification analysis of Fiverr Application user ratings on Google Play sites using the Naïve Bayes Classifier method. Data collection on Fiverr App reviews uses web scraping techniques through the Google Collab website. The data that has been obtained is then labeled between positive, negative, or neutral. After being labeled, text preprocessing and TF-IDF weighting are carried out in each review. Furthermore, the classification uses the Naïve Bayes model with 454 data training and 454 data testing. The classification results show that Fiver App reviews a total of 454 data tests, showing a percentage of accuracy of 85,24%, precision of 97,59%, and recall of 88.20%