ANALISIS SENTIMEN PADA TWITTER MENGGUNAKAN METODE MAXIMUM ENTROPY DAN SUPPORT VECTOR MACHINE
Daftar Isi:
- Social media such as Twitter are often used by users to give opinions about the quality of products they use, express their emotions every day and some companies such as airlines use to gain competitive advantage by continuing to improve service and experience of service users by using twitter to become an important source for tracking sentiment analysis. Sentiment analysis is done to see opinions or tendencies of opinion on a problem. This study uses two types of datasets, namely public opinion data and airline opinion data to test the best features and methods. Feature extraction is expected to improve the accuracy of sentiment analysis. To get the best results this research uses two methods, namely maximum entropy and support vector machine as a comparison. The best results are found in the maximum entropy method which has an average accuracy of 85.8 % for the general opinion dataset while in the airline's opinion dataset the accuracy reaches 92.6%.