Public services by the government generally have an impact that is quickly responded to by the community. One form of public response is through their opinions through writings written on social media or reviews of applications developed by the government. Machine learning has been widely used for automatic opinion mining to classify sentiment classes. The classification method that can be used to classify public opinion into positive or negative sentiment classes is random forest. Based on the test results of the random forest algorithm in classifying sentiments from user reviews of public service applications by the government, the highest accuracy value was obtained at 84% by performing hyperparameter tuning.
Copyrights © 2022