Analisis Sentimen Di Media Sosial X Terhadap Event Pemilihan Umum Di Indonesia Menggunakan Metode Bi-Lstm Dan Bi-Gru

Herprasetyo, Handitanto (2024) Analisis Sentimen Di Media Sosial X Terhadap Event Pemilihan Umum Di Indonesia Menggunakan Metode Bi-Lstm Dan Bi-Gru. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5025201077_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (4MB)

Abstract

Sosial media telah menjadi platform utama untuk pertukaran informasi di era digital,
memungkinkan penyebaran cepat informasi yang akurat maupun tidak terverifikasi. Platform
seperti X (Twitter), Facebook, dan Instagram telah menjadi bagian integral dari kehidupan
sehari-hari, digunakan oleh individu dan institusi untuk berkomunikasi dan mempengaruhi
opini publik. Dalam konteks pemilu, media sosial, terutama X, memainkan peran penting dalam
diskusi politik dan mobilisasi massa, seperti yang terlihat dalam pemilu Indonesia yang
dilakukan setiap lima tahun sekali. Dinamika ini menunjukkan bahwa analisis sentimen di
media sosial menjadi kunci untuk memahami respons dan pandangan masyarakat terhadap
calon, partai politik, dan isu-isu terkait pemilu. Dengan teknik analisis bahasa alami, analisis
sentimen membantu mengukur perasaan kolektif masyarakat, mendeteksi tren pemilih, dan
merancang strategi kampanye yang efektif, sehingga meningkatkan transparansi, akuntabilitas,
dan integritas dalam proses demokrasi.
Penelitian melibatkan pengumpulan dan pra-pemrosesan data teks, termasuk case
folding, penghapusan tautan, dan lainnya. Setelah itu, data dilabeli secara manual dan dengan
model pra-terlatih IndoBERT untuk klasifikasi ke dalam label positif, negatif, dan netral.
Parameter tuning diterapkan untuk menghasilkan model Bi-LSTM dengan akurasi 0.9050,
precision 0.9049, dan F1-score 0.9043. Analisis topik dengan metode LDA menunjukkan
dominasi sentimen negatif pada topik seperti "Pemilihan Legislatif", "Pemilihan Presiden", dan
"Sistem dan Perhitungan Suara dalam Pemilu".
Analisis keterkaitan antara sentimen masyarakat di media sosial X dengan berita-berita
populer terkait event pemilu menunjukkan bahwa berita-berita populer yang ditemukan
memang sejalan dengan sentimen negatif yang mendominasi dataset tweet. Isu-isu yang terkait
dengan pelanggaran hukum oleh pejabat, ketidakpuasan terhadap kinerja politik, dan
ketidakpercayaan terhadap proses pemilu secara keseluruhan tampaknya sangat mempengaruhi
persepsi dan reaksi negatif masyarakat di platform X.

============================================================

Social media has become a primary platform for information exchange in the digital era,
enabling the rapid spread of both accurate and unverified information. Platforms such as X
(Twitter), Facebook, and Instagram have become integral parts of daily life, used by individuals
and institutions to communicate and influence public opinion. In the context of elections, social
media, especially X, plays a crucial role in political discussions and mass mobilization, as seen
in Indonesia's elections held every five years. This dynamic indicates that sentiment analysis
on social media is key to understanding public responses and opinions towards candidates,
political parties, and election-related issues. By using natural language processing techniques,
sentiment analysis helps measure the collective feelings of the public, detect voter trends, and
design effective campaign strategies, thereby enhancing transparency, accountability, and
integrity in the democratic process.
The research involves the collection and preprocessing of text data, including case
folding, link removal, and more. After that, the data is labeled manually and with a pre-trained
IndoBERT model for classification into positive, negative, and neutral labels. Parameter tuning
was applied to produce a Bi-LSTM model with an accuracy of 0.9050, precision of 0.9049, and
an F1-score of 0.9043. Topic analysis using the LDA method shows a dominance of negative
sentiment on topics such as "Legislative Elections," "Presidential Elections," and "Election
Vote Counting Systems."
The analysis of the relationship between public sentiment on X social media and popular
news related to election events shows that the popular news found aligns with the negative
sentiment dominating the tweet dataset. Issues related to legal violations by officials,
dissatisfaction with political performance, and distrust in the overall election process appear to
significantly influence public perception and negative reactions on the X platform.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Analisis Sentimen, Media Sosial X, Pemilihan Umum, Bi-LSTM, Bi-GRU, Sentiment Analysis, X Social Media, General Elections, Bi-LSTM, Bi-GRU
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Handitanto Herprasetyo
Date Deposited:	07 Aug 2024 08:17
Last Modified:	07 Aug 2024 08:17
URI:	http://repository.its.ac.id/id/eprint/111753

Actions (login required)

View Item