ARYANI, DINI (2024) PEMODELAN TOPIK PEMILU 2024 MENGGUNAKAN METODE BERTOPIC PADA ARTIKEL BERITA DETIK.COM. Other thesis, Nusa Putra University.
DINI ARYANI .pdf
Download (978kB)
Abstract
In 2024, Indonesia held simultaneous general elections dominated by the participation of young generations, especially Generation Z and millennials. This generation more often seeks political information through the internet, highlighting the important role of digital media in shaping public opinion. Research shows that positive exposure to digital media can increase support and political participation among the millennial generation. This study uses Topic Modelling techniques to analyze topics related to the 2024 election from news articles on Kompas.com, using the BERTopic method. The dataset comes from the Detik.com news portal in the ''pemilu'' sub-channel, collected through scraping from September 1, 2023, to February 14, 2024, totaling 15,019 articles. The text preprocessing process involves four stages: case folding, cleaning, tokenizing, and stopword removal. Topic modeling uses BERTopic through five stages: embeddings, dimensionality reduction, clustering, tokenizer, and weighting scheme. The modeling results show a topic coherence value of 0.0885752. This study successfully identified five main topics in the news coverage of the 2024 ''pemilu'' on the Detik.com news portal. These topics include: Topic 0, focusing on ''calon-calon presiden'' and ''wakil presiden'' reflecting high attention to key figures in the ''pemilu'', with a total of 6211 news articles with representative words 'ganjar', 'prabowo', 'anies', 'imin'; Topic 1, focusing on ''nomor urut'' of ''pasangan capres- cawapres'' in the general election in Indonesia, particularly mentioning figures such as ''Gibran Rakabuming Raka'', with a total of 2661 news articles with representative words 'nomor', 'urut', 'gibran', 'indonesia'; Topic 2, about news on ''Presiden Joko Widodo'' and his political dynamics, with a total of 2615 news articles with representative words 'presiden', 'politik', 'jokowi', 'survei'; Topic 3, related to the general election and ''pemilihan presiden'', including votes and related surveys, with 2349 news articles with representative words '2024', 'pemilu', 'pilpres', 'kpu'; Topic 4, about news on ''debat capres'' and ''cawapres'' that occurred, with a total of 1183 news articles with representative words 'debat', 'pilpres', 'cawapres', '2024'.
Keywords: Topic modelling, BERTopic, 2024 election, news, Detik.com
| Item Type: | Thesis (Other) |
|---|---|
| Subjects: | Computer > Informatic Engineering |
| Divisions: | Faculty of Engineering, Computer and Design > Informatic Engineering |
| Depositing User: | Mr Perpus |
| Date Deposited: | 19 Jan 2025 07:03 |
| Last Modified: | 19 Jan 2025 07:03 |
| URI: | http://repository.nusaputra.ac.id/id/eprint/1300 |
