| Summary: | Twitter now is the one of the most influental social media in the world, especially in Indonesia. Unfortunately, research about sentiment analysis in Bahasa (Indonesian Language) are not that many although there are many possibilities to convert the opinion inside into valuable information. One of the kind of information that can be made is rate. This research is aim to do sentiment analysis in Twitter with Bahasa so that the opinions can be converted to a rate.
There are two steps in this ressearch. The first one is building the core of the application using lucene library and support vector machine. The second one is to build the application using Java.
The result is, the core of the application�s accuracy is 74.34%, while the application instead give rate 4.4 for keyword �Jokowi� with 68% of accuracy, rate 2.7 for keyword �Prabowo� with 56% of accuracy, rate 4.6 for keyword �kalimilk� with 70% of accuracy and it also gave rate 3.7 for keyword �sunmor� with 74% of accuracy.
|
|---|