Penambangan Topik Pada Teks Berita Berbahasa Indonesia Menggunakan Dynamic Topic Modeling

Adibah, Aaliyah Farah (2024) Penambangan Topik Pada Teks Berita Berbahasa Indonesia Menggunakan Dynamic Topic Modeling. Other thesis, lnstitut Teknologi Sepuluh Nopember.

Text 5025201070-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (22MB) | Request a copy

Abstract

Cepatnya perkembangan internet di Indonesia membawa pengaruh yang besar bagi kehidupan masyarakat. Salah satu pengaruh yang terasa adalah masyarakat Indonesia memiliki akses yang lebih mudah dan cepat untuk mendapat berita terbaru melalui platform online. Banyaknya pertambahan berita terbaru setiap saat dapat menimbulkan isu bagi pengguna dalam mendapatkan informasi mengenai topik berita yang sedang dibahas. Dalam konteks ini, penting untuk mengatasi tantangan informasi yang berlimpah dengan penambangan topik. Penambangan topik dilakukan menggunakan pemodelan topik sehingga mempermudah masyarakat dalam topik berita yang sedang dibahas, terlebih lagi apabila terbagi dalam sekuensial waktu tertentu. Berdasarkan kebutuhan tersebut, dilakukannya pengembangan pemodelan topik berdasarkan sekuensial waktu pada teks berita berbahasa Indonesia yang dinamakan dynamic topic modeling. Penelitian ini akan berfokus untuk membandingkan kedua model dynamic topic modeling, yaitu Hierarchical Dirichlet Process (HDP) dan BERTopic terhadap implementasinya pada teks berita berbahasa Indonesia. Berdasarkan uji skenario yang dilakukan di penelitian, model BERTopic memberikan performa terbaik dalam penambangan topik dinamis dengan coherence score sebesar 0,81112. Melalui penelitian ini, diharapkan penelitian ini dapat memberikan kontribusi dalam pengembangan metode penambangan topik berdasarkan irisan waktu sehingga pembaca bisa mengetahui topik yang sedang banyak dibahas.
============================================================================================================================
The rapid development of the internet in Indonesia has had a huge impact on people's lives. One of the most noticeable influences is that Indonesians have easier and faster access to the latest news through online platforms. The continuous increase of new news can cause issues for users in obtaining information about the current news topics being discussed. In this context, it is necessary to overcome the challenges of abundant information with topic mining. Topic mining is performed using topic modeling to make it easier for people to follow the current news topics being discussed, especially if they are divided into certain time sequences. Based on these needs, this research will develop topic modeling based on time sequence in Indonesian news text called dynamic topic modeling. This research will focus on comparing the two dynamic topic modeling models, namely Hierarchical Dirichlet Process (HDP) and BERTopic towards their implementation on Indonesian news texts. Based on the scenario test conducted in the research, the BERTopic model provides the best performance in dynamic topic mining with a coherence score of 0.81112. Through this research, it is hoped that this research can contribute to the development of topic mining methods based on time slices so that readers can find out the current news topics being discussed.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	News, Berita, Dynamic Topic Modeling, BERTopic, HDP
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science)
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Aaliyah Farah Adibah
Date Deposited:	01 Aug 2024 15:06
Last Modified:	01 Aug 2024 15:06
URI:	http://repository.its.ac.id/id/eprint/111585

Actions (login required)

View Item