PENGGUNAAN TOPIC MODELING DENGAN METODE LATENT DIRICHLET ALLOCATION TERHADAP BERITA SAHAM INDEKS LQ45

Anwar, Akmal Aliffandhi (2025) PENGGUNAAN TOPIC MODELING DENGAN METODE LATENT DIRICHLET ALLOCATION TERHADAP BERITA SAHAM INDEKS LQ45. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 20081010148-cover.pdf Download (978kB)
	Text (Bab 1) 20081010148-bab 1.pdf Download (145kB)
	Text (Bab 2) 20081010148-bab 2.pdf Restricted to Repository staff only until 6 May 2026. Download (319kB)
	Text (Bab 3) 20081010148-bab 3.pdf Restricted to Repository staff only until 6 May 2026. Download (433kB)
	Text (Bab 4) 20081010148-bab 4.pdf Restricted to Repository staff only until 6 May 2026. Download (1MB)
	Text (Bab 5) 20081010148-bab 5.pdf Download (79kB)
	Text (Daftar Pustaka) 20081010148-daftar pustaka.pdf Download (77kB)

Abstract

Stock investment in Indonesia has become an attractive choice due to its high potential returns despite significant risks. The COVID-19 pandemic impacted the decline of the Jakarta Composite Index (JCI) but increased the number of investors, especially millennials and Gen Z, driven by the ease of technology in stock transactions. Information analysis is crucial for investment decision-making, but challenges arise from the growing volume and complexity of news. Topic modeling techniques such as Latent dirichlet allocation (LDA) are used to address these challenges, enabling automatic and efficient identification of main themes. This study applies LDA to LQ45 stock news through data preprocessing, training and testing dataset splitting, and coherence score evaluation. The results reveal five main topics, including "Technical Stock Analysis" and "Bank Financial Performance," with the highest coherence score of 0.66238 on five topics. Although LDA performs lower than HDP and NMF, this method effectively accelerates analysis and provides a structured understanding of stock market dynamics.

Item Type:

Thesis (Undergraduate)

Contributors: