OPINION MINING PADA TWITTER UNTUK BAHASA INDONESIA DENGAN METODE SUPPORT VECTOR MACHINE DAN METODE BERBASIS LEXICON; OPINION MINING ON TWITTER FOR INDONESIAN USING SUPPORT VECTOR MACHINE AND LEXICON BASED METHOD

JAN KRISTANTO WIBISONO, Edi Winarko

2013 | Disertasi | PROGRAM STUDI S2 ILMU KOMPUTER

Opinions in the decision-making process are often used as a reference. Nowadays, with the internet and social media there are a lot of people often argue, complain or criticize about something. So as to know what others think of something that we do not need to ask them one by one. We can develop a method that can search and analyze text on social media which produces output in the form an opinion about something. The method to obtain information in the form of opinions contained in the text is an "Opinion Mining" or "Sentiment Analysis". In this research, using a combined Unsupervised method based on Indonesian Lexicon and Supervised based on Support Vector Machine. Results obtained in the form of sentiment is negative polarity, positive, or neutral. From the test results obtained in this study 77,7% accuracy.

Kata Kunci : opinion mining; sentiment analysis; lexicon; support vector machine