Topic Modeling in Thesis Titles of Students from the Faculty of Economics Universitas Garut Using Latent Dirichlet Allocation Modeling

Abstract

In higher education, the completion of a thesis within a 1 -year timeframe is a prerequisite for graduation. The selection of a thesistopic is influenced by personal interest, the expertiseof the thesis supervisor, and data availability. This research is designed to analyzethe thesis topics of Economics Faculty students at Garut University using the Latent Dirichlet Allocation (LDA) Modeling method. Utilizing quantitative and qualitative approaches, this research applies the concept of big data with techniques such as Data Crawling, Data Preprocessing, and Text Mining. The research successfully conducted topic modeling using the LDA method. The analysis showed that topic modeling with the LDA algorithm resulted in seven common thesis topics used in the students' thesis titles. With this, theresearch contributes to the understanding and efficacy in the determination of students' thesis topics. It is hoped that the results of this research can be utilized to assist in the efficient completion of theses.