Peningkatan Hasil Analisa Sentimen Menggunakan Pos Tagger Untuk Melihat Tanggapan Masyarakat Terhadap Full Day School

Nowadays development in information technology is continually increasing. Almost all information can be obtained easily through the internet. Information access can be obtained not only through the online news but also through social networking media such as Facebook, Twitter or Instagram. Such information can be used for specific purposes such as determining the value of trust in the online shop, online transaction extracting, assessment of public figures and determine the community assessment of government policies, such as full day school. The government's policy that will be made caused people who is agree and people who is disagree. It is becoming a problem because majority of people who is agree or not can not be known. This problem will be investigated using a lexicon based approach because the sentiment value will be calculated word by word in each sentence and, the process is fast. Lexion based approach would be assisted by the library of Stanford POS Tagger to improve the observation results. Calculation which produced by the application is 98 positive sentiments, 90 negative sentiments and 27 neutral sentiments. The result show that the people agree with the the full day school program. This research provides an increasing 0,042 of accuracy obtained from comparison of application with POS Tagger and application without POS Tagger