Advances in technology make people tend to choose online transactions compared to cash. With the rapid development of technology, PT Pertamina launched their newest program, the MyPertamina application. The MyPertamina application is a combination program between loyalty and e-payment. The MyPertamina application has many users, so there are often positive and negative reviews that are irrelevant to the rating given on the Google Play Store. MyPertamina application review data will be obtained using the Web Scraping technique using the Google-Play-Scraper API. The scraped data will then be subjected to text preprocessing to clean the text so that the data can be executed. Sentiment analysis can detect whether a text contains positive or negative opinions from user reviews that are not in accordance with the rating. Random Forest is a method in analysis consisting of several decision trees as a classifier. In this study, the random forest method was used by dividing two sentiment classes, namely positive and negative, taking evaluation indicators, namely accuracy, recall, precision and f1-score. Tests were carried out based on the number of trees and tree depth on 800 data by dividing the data by 250 data per class. Based on the results of the tests and analyzes that have been carried out, with a comparison of training data and test data 80%:20% obtained values of 90% accuracy, 90% precision, 91% recall, and 89% F1-Score with a depth of 80 trees and 50 tree depths.
Copyrights © 2023