COMPARISON OF LSA AND LDA ALGORITHMS IN TOPIC MODELING

Aris Subadi; Endang Kusnadi

COMPARISON OF LSA AND LDA ALGORITHMS IN TOPIC MODELING

Authors

Aris Subadi STKIP Mutiara Banten
Endang Kusnadi

Keywords:

Topic Modelling, Latent Dirichlet Allocation (LDA), Latent Semantic Analysis (LSA), Tiket Insiden , Data analysist

Abstract

This study aims to analyze topic modeling in incident ticket reports by comparing two methods, namely Latent Semantic Analysis (LSA) and Latent Dirichlet Allocation (LDA). The topic modeling method is used to uncover hidden topics in unstructured text. This study involved the initial stages of data preprocessing to convert unstructured text into structured text by changing the letters to lowercase, removing distracting characters, and removing stopwords. After the preprocessing stage was completed, topic modeling was carried out using LSA and LDA with variations in the number of topics. Performance evaluation was carried out using coherence scores and t-SNE visualization for each method. The evaluation results show that LSA has a higher coherence score for several topics, especially for topics 3 and 6, with a coherence score of around 0.57. On the other hand, LDA provides a better interpretation of the distribution of words in each topic, although the coherence score is slightly lower. The LDA coherence score is around 0.53 for topic 3, 0.43 for topic 6, and 0.46 for topic 10. In addition, a comparison between LDA and LSA can also assist researchers in choosing the most suitable method for topic analysis on a particular dataset.

References

[1] I. A. Amala, D. Richasdy, and M. D. Purbolaksono, “Telkom University News Topic Modeling Using Latent Semantic Analysis (LSA) Method on Online News Portal,” Building of Informatics, Technology and Science (BITS), vol. 4, no. 1, Jun. 2022, doi: 10.47065/bits.v4i1.1584.

[2] S. Karmila, V. Intan Ardianti Prodi Telematika Energi, and V. Intan Ardianti, “METODE LATENT DIRICHLET ALLOCATION UNTUK MENENTUKAN TOPIK TEKS SUATU BERITA,” Jurnal Informatika & Komputasi, vol. 16, 2022.

[3] S. H. Mohammed and S. Al-Augby, “LSA & LDA topic modeling classification: Comparison study on E-books,” Indonesian Journal of Electrical Engineering and Computer Science, vol. 19, no. 1, pp. 353–362, 2020, doi: 10.11591/ijeecs.v19.i1.pp353-362.

[4] S. Bellaouar, M. M. Bellaouar, and I. E. Ghada, “Topic modeling: Comparison of LSA and LDA on scientific publications,” in ACM International Conference Proceeding Series, Association for Computing Machinery, Feb. 2021, pp. 59–64. doi: 10.1145/3456146.3456156.

[5] D. Rolliawati, U. Sunan, and A. Surabaya, “Topic Modeling Pada Abstrak Skripsi Menggunakan Metode Latent Semantic Analysis,” 2022, doi: 10.22441/format.2022.v11.i1.009.

[6] Vaigai College of Engineering and Institute of Electrical and Electronics Engineers, Proceedings of the International Conference on Intelligent Computing and Control Systems (ICICCS 2020) : 13-15 May, 2020.

[7] E. S. Negara, D. Triadi, and R. Andryani, “Topic Modelling Twitter Data with Latent Dirichlet Allocation Method.”

[8] S. Hassan, M. Al-Augby, S. H. Mohammed, and S. Al-Augby, “LSA & LDA Topic Modeling Classification: Comparison study on E-books Sentiment Analysis on Social Media View project LSA & LDA Topic Modeling Classification: Comparison study on E-books,” Indonesian Journal of Electrical Engineering and Computer Science, vol. 19, no. 1, 2020, doi: 10.11591/ijeecs.v19.i1.pp%25p.

[9] S. Devi, C. Dhavale, L. Moharkar, and S. Khanvilkar, “Impact of Online Education and Sentiment Analysis from Twitter Data using Topic Modeling Algorithms,” International Journal of Applied Sciences and Smart Technologies, vol. 4, no. 1, p. 21.

[10] A. Nurlayli and M. A. Nasichuddin, “Topic Modeling Penelitian Dosen JPTEI UNY pada Google Scholar Menggunakan Latent Dirichlet Allocation,” vol. 4, no. 2, pp. 154–161, 2019, doi: 10.21831/elinvo.v4i2.

[11] U. T. Setijohatmo et al., “Prosiding The 11 th Industrial Research Workshop and National Seminar Bandung,” 2020.

[12] T. Williams and J. Betak, “A Comparison of LSA and LDA for the Analysis of Railroad Accide Text,” in Procedia Computer Science, Elsevier B.V., 2018, pp. 98–102. doi: 10.1016/j.procs.2018.04.017.

Downloads

Published

2026-07-20

How to Cite

Subadi, A., & Kusnadi, E. (2026). COMPARISON OF LSA AND LDA ALGORITHMS IN TOPIC MODELING. Journal of Civil Engineering, Technology and Sciences, 2(2), 74–87. Retrieved from https://jcets.journaldpupr.info/index.php/jocets/article/view/76