Debiasing Pemberitaan Online terhadap Berita Pemilu Presiden di Indonesia

Yuniarti, Lidiya (2024) Debiasing Pemberitaan Online terhadap Berita Pemilu Presiden di Indonesia. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6026231027-Master_Thesis.pdf Download (3MB)

Abstract

Pemberitaan politik mengenai presiden dalam media massa termasuk berita online
yang memiliki peran penting dalam proses pemilu dan dalam membentuk opini
publik. Namun, pemberitaan tersebut cenderung rentang terhadap bias (seperti
nama kandidiat, partai politik dan organisasi yang terlibat dalam pemilu) sehingga
dapat memengaruhi persepsi dan partisipasi politik dari masyarakat. oleh karena
itu, penting untuk mengindentifikasi dan mengatasi bias dalam pemberitaan
politik khususnya terkait pemilu presiden di Indonesia. Adapun metodologi yang
digunakan dalam penelitian ini menggunakan word embedding yaitu model
Word2Vec dan IndoBERT dalam menganalisis resepsentasi dalam pemberitaaan
pada berita politik khususnya pemilu presiden di Indonesia dan menggunakan
sentimen analisis berbasis leksikon untuk memastikan proses debiasing yang
diterapkan berhasil mengurangi bias. Adapun hasil dari penelitian ini representasi
bias menggunakan word embedding berhasil dilakukan, namun metode sentimen
analisis berbasis leksikon yang digunakan tidak efektif dalam melakukan
melakukan evaluasi bias pada berita pemilu presiden 2024 di Indonesia dibuktikan
dengan naiknya nilai sentimen yang menunjukkan kecenederungan sentimen
dalam analisis seperti naiknya sentimen negatif sebesar 0,18 dan sentimen positif
sebesar 2,05 serta terjadi penurunan nilai sentimen netral sebesar 2,24 setelah
dilakukan proses debiasing.
========================================================================================================================
Political news coverage regarding the president in mass media, including online
news, plays an important role in the election process and in shaping public
opinion. However, such coverage tends to be susceptible to bias (such as the
names of candidates, political parties, and organizations involved in the election),
which can influence public perception and political participation. Therefore, it is
important to identify and address bias in political news coverage, especially
related to the presidential election in Indonesia. The methodology used in this
study involves word embeddings, specifically the Word2Vec and IndoBERT
models, to analyze representations in political news, particularly the Indonesian
presidential election. Additionally, a lexicon-based sentiment analysis is used to
ensure that the debiasing process effectively reduces bias. The findings of this
study indicate that bias representation using word embeddings was successfully
conducted. However, the lexicon-based sentiment analysis method used was not
effective in evaluating bias in the news of the 2024 Indonesian presidential
election. This is evidenced by the increase in sentiment values, indicating a
tendency in sentiment analysis, with a rise in negative sentiment by 0.18, positive
sentiment by 2.05, and a decrease in neutral sentiment by 2.24 after the debiasing
process.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Debiasing, isu calon presiden, word embedding, sentimen analisis, Debiasing, presidential candidate issues, word embedding, sentiment analysis .
Subjects:	Z Bibliography. Library Science. Information Resources > ZA Information resources > Z699.5 Information storage and retrieval systems
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 59101-(S2) Master Thesis
Depositing User:	Lidiya Yuniarti
Date Deposited:	07 Aug 2024 22:51
Last Modified:	07 Aug 2024 22:51
URI:	http://repository.its.ac.id/id/eprint/113217

Actions (login required)

View Item