Analisis sentimen terkait isu Agama menjelang pemilihan umum 2019 Indonesia menggunakan Deep Belief Network

Susilowati, Qoriah Indah (2019) Analisis sentimen terkait isu Agama menjelang pemilihan umum 2019 Indonesia menggunakan Deep Belief Network. Diploma thesis, UIN Sunan Gunung Djati Bandung.

Preview	Text (COVER) 1_Cover.pdf Download (611kB) | Preview
Preview	Text (ABSTRAK) 2_Abstrak.pdf Download (654kB) | Preview
Preview	Text (DAFTAR ISI) 3_Daftar Isi.pdf Download (599kB) | Preview
Preview	Text (BAB I) 4_Bab1.pdf Download (913kB) | Preview
	Text (BAB II) 5_Bab2.pdf Restricted to Registered users only Download (977kB) | Request a copy
	Text (BAB III) 6_Bab3.pdf Restricted to Registered users only Download (1MB) | Request a copy
	Text (BAB IV) 7_Bab4.pdf Restricted to Registered users only Download (1MB) | Request a copy
	Text (BAB V) 8_Bab5.pdf Restricted to Registered users only Download (654kB) | Request a copy
	Text (DAFTAR PUSTAKA) 9_Daftar Pustaka.pdf Restricted to Registered users only Download (659kB) | Request a copy

Abstract

General election in 2019 is one of democracy system for electing president and his vice mediapresindent and electing citizen’s representatives to sit in parliement. General election take more attention for social media users, especially twitter. It was supported by most of Indonesia political figures who have twitter account to provide support or theirs respective pair of candidates. Among various aspects, religion is the most used issues through election. Religious sentiments or opinions used are very varied contain neutral, positive even not a few that give negative sentiments for presindent candidates. This research was conducted to determine the optimization of Deep Belief Network (DBN) algorithm to classify sentiments on religious issues used through of the 2019 Indonesia elections with data obtained from Twitter, then preprocessing and weighting using TF-IDF. DBN algorithm is tested 30 times with different epochs and hidden layers to find optimal results. Testing with hidden layer 10 produces an accuracy of 52.46% and testing with an increase in the number of epochs and hidden layer produces an accuracy of 52.53%. This research can get that the 2019 general election gives a lot of sentiment related to religious issues in the time leading up to the election. This is evidenced from the fact that 65% of the data is positive sentiment containing religious issues.

Actions (login required)

View Item