Comparing LDA and BERTopic for Topic Modeling of Skin Complaints in Indonesian-Language Sunscreen Reviews on Shopee

Fahmi Ilmiah; Arif  Faizin

doi:10.37899/journallamultiapp.v7i4.5185

Fahmi Ilmiah Department of Informatic Engineering, Faculty of Engineering, Universitas Yudharta Pasuruan, Indonesi
Arif Faizin Department of Informatic Engineering, Faculty of Engineering, Universitas Yudharta Pasuruan, Indonesi

DOI: https://doi.org/10.37899/journallamultiapp.v7i4.5185

Keywords: Sunscreen, Topic Modeling, LDA, BERTopic, Tropical Climate

Abstract

Sunscreen use in Indonesia continues to grow alongside the high ultraviolet exposure typical of a tropical climate. However, various complaints such as a sticky texture, white cast, and the appearance of acne are still frequently found in user reviews. These reviews contain important information about users' experiences and skin problems, yet they are difficult to analyze manually because of their large volume and unstructured nature. This study therefore aims to identify skin-complaint patterns in sunscreen use through an analysis of user reviews on Shopee Indonesia. It applies a text mining approach using topic modeling methods, namely Latent Dirichlet Allocation (LDA) and BERTopic. Sunscreen product reviews were collected from the Shopee platform and then passed through preprocessing stages comprising cleaning, case folding, tokenizing, normalization, and stemming. The two methods were subsequently used to extract topics representing user complaints. Evaluation was carried out using the coherence score, topic diversity, and manual interpretation to assess the quality of the resulting topics. Based on the results, LDA produced nine topics with a coherence score of 0.35 and a topic diversity of 0.70, whereas BERTopic produced twelve topics with a coherence score of 0.50 and a topic diversity of 0.90. Based on the evaluation and an examination of dominant words and representative reviews, the topics produced by BERTopic showed better interpretability because they were more specific and easier to label, whereas LDA produced more general topics that still combined several complaints within a single topic. Accordingly,

References

Aban, N. A., & Ratnasari, C. I. (2026). Pemodelan Topik Cuitan Tentang Danantara Menggunakan Bertopic Teroptimasi Umap Dan Hdbscan. Rabit : Jurnal Teknologi Dan Sistem Informasi Univrab, 11(1), 1160–1174. https://doi.org/10.36341/rabit.v11i1.7238

Allam, H., Makubvure, L., Gyamfi, B., Graham, K. N., & Akinwolere, K. (2025). Text classification: How machine learning is revolutionizing text categorization. Information, 16(2), 130. https://doi.org/10.3390/info16020130

Anjarsari, D. S., Imaningsih, A. N., Arviani, I., & Prasetyo, A. Y. (2025). Hubungan Penggunaan Face Mist Berbahan Dasar Alami Terhadap Kesehatan Kulit Wajah pada Cuaca Ekstrem Di kalangan Masyarakat Kabupaten Jombang. Journal Of Social Science Research, 5, 1090–1102.

Arigi, J. P., Hidayat, Z., Putra, D. O., & Gulo, A. (2024). Analysis of metrosexual consumer lifestyle in urban areas: A phenomenological study of male-targeted beauty consumers on social media. Humanities, Arts and Social Sciences Studies, 768-780. https://doi.org/10.69598/hasss.24.3.264857

Asnawi, M. F., Hanafi, M., Kurniawan, N. F., Suwondo, A., Nasrullah, A., & Setyawan, C. (2024, November). Topic modelling analysis on indonesian news using bert topic model. In 2024 6th International Conference on Cybernetics and Intelligent System (ICORIS) (pp. 1-6). IEEE. https://doi.org/10.1109/ICORIS63540.2024.10903779

Beauté, R., Schwartzman, D. J., Dumas, G., Crook, J., Macpherson, F., Barrett, A. B., & Seth, A. K. (2026). Mapping of Subjective Accounts into Interpreted Clusters (MOSAIC): topic modelling and LLM applied to stroboscopic phenomenology. Neuroscience of Consciousness, 2026(1), niag008. https://doi.org/10.1093/nc/niag008

Chen, Y., Peng, Z., Kim, S. H., & Choi, C. W. (2023). What we can do and cannot do with topic modeling: A systematic review. Communication Methods and Measures, 17(2), 111-130. https://doi.org/10.1080/19312458.2023.2167965

Chu, K. E., Keikhosrokiani, P., & Asl, M. P. (2022). A topic modeling and sentiment analysis model for detection and visualization of themes in literary texts. Pertanika Journal of Science & Technology, 30(4), 2535-2561.

Councill, I. G., Mcdonald, R., & Velikovich, L. (2010). What ’ s Great and What ’ s Not : Learning to Classify the Scope of Negation for Improved Sentiment Analysis. July, 51–59.

Demoze, L., Gubena, F., Akalewold, E., Brhan, H., Adane, K. C., Kifle, T., ... & Yitageasu, G. (2024). Burden and determinants of scabies in Ethiopian school age children: A systematic review and meta-analysis with public health implications. PloS one, 19(12), e0314882. https://doi.org/10.1371/journal.pone.0314882

Dieng, A. B., Ruiz, F. J. R., & Blei, D. M. (2020). Topic modeling in embedding spaces. Transactions of the Association for Computational Linguistics, 8, 439–453. https://doi.org/10.1162/tacl_a_00325

Fitraneti, E., Rizal, Y., Riska Nafiah, S., Primawati, I., & Ayu Hamama, D. (2024). Pengaruh Paparan Sinar Ultraviolet terhadap Kesehatan Kulit dan Upaya Pencegahannya : Tinjauan Literatur. Scientific Journal, 3(3), 185–194. https://doi.org/10.56260/sciena.v3i3.147

Grassia, M. G., Marino, M., Mazza, R., Misuraca, M., Zavarrone, E., & Friel, M. (2024). Regional Competitiveness: A Structural-Based Topic Analysis on Recent Literature: MG Grassia et al. Social Indicators Research, 173(1), 83-108. https://doi.org/10.1007/s11205-022-02951-4

Grootendorst, M. (2020). BERTopic: Neural topic modeling with a class-based TF-IDF procedure.

Guenduez, A. A., Fuchs, S., Mergel, I., Mettler, T., & Frowein, S. (2026). From patterns to meaning: a mixed-methods framework that integrates computational and qualitative text analysis. Quality & quantity, 60(1), 2713-2738. https://doi.org/10.1007/s11135-025-02357-7

Harahap, D. K., Tania, K. D., & Sevtiyuni, P. E. (2025). Topic Mining-Based Knowledge Discovery of User Health Information Needs. Indonesian Journal of Electronics, Electromedical Engineering, and Medical Informatics, 7(4), 641-653. https://doi.org/10.35882/ijeeemi.v7i4.270

Huang, Z., & Jiang, Z. (2026). Text mining of syntactic complexity in L2 writing: An LDA topic modeling approach. International Review of Applied Linguistics in Language Teaching, 64(1), 523-548.

Jung, I. S., Yun, S. J., & Lee, J. B. (2019). The Difference in Sebum Secretion Affecting Development of Acne. Annals of Dermatology, 31(4), 426–433.

Li, D., Wu, K., & Lei, V. L. (2024). Applying topic modeling to literary analysis: A review. Digital Studies in Language and Literature, 1(1-2), 113-141.

Mertayasa, I. K. T., & Wulandhari, L. A. (2025, November). Topic Modeling of Indonesian-Language Travel Reviews on Things to Do in Bali Using BERTopic and IndoBERT. In 2025 13th International Conference on Awareness Science and Technology (iCAST) (pp. 1-7). IEEE. https://doi.org/10.1109/iCAST68191.2025.11300049

Moloney, F. J., Collins, S., & Murphy, G. M. (2002). Sunscreens: safety, efficacy and appropriate use. American journal of clinical dermatology, 3(3), 185-191. https://doi.org/10.2165/00128071-200203030-00005

Nabila, N., & Ratnasari, C. I. (2025). Topic Modeling of Skincare Comments from Female Daily. Journal of Applied Informatics and Computing, 9(4), 1394–1405. https://doi.org/10.30871/jaic.v9i4.9625

Nurcholis, J. D., Hadibrata, M., Arzakiyah, M. M., Nasution, F. B., & Yuliandari, P. (2026). Analisis Big Data pada Perdagangan Digital: Studi Komparatif E-Commerce, S-Commerce, dan Q-Commerce Berbasis BERTopic. INOMATEC: Jurnal Inovasi dan Kajian Multidisipliner Kontemporer, 1(3), 417–431.

Nursyahrina, Defit, S., & Sovia, R. (2024). Metode BERTopic dan LDA untuk Analisis Tren Penelitian Bidang Ilmu Komputer. Jurnal KomtekInfo, 11(4), 332–341. https://doi.org/10.35134/komtekinfo.v11i4.580

Parasdita, A. C., & Kurniasari, F. (2025). The Influence of Promotion , Online Customer Review , and Online Customer Rating on Purchasing Decisions in the Shopee Application Among Students in Pontianak City. 6(2), 1–9.

Puspita, E., Shiddieq, D. F., & Roji, F. F. (2024). Pemodelan Topik pada Media Berita Online Menggunakan Latent Dirichlet Allocation (Studi Kasus Merek Somethinc). MALCOM: Indonesian Journal of Machine Learning and Computer Science, 4(2), 481–489. https://doi.org/10.57152/malcom.v4i2.1204

Putri, D. H., Santosa, I., Junaidy, D. W., & Adhitama, G. P. (2026). A Systematic Literature Review of Spatial Attributes and Well-Being in Women’s Spa Experiences. The International Journal of Architectonic, Spatial, and Environmental Design, 20(1), 215.

Rüdiger, M., Antons, D., Joshi, A. M., & Salge, T. O. (2022). Topic modeling revisited: New evidence on algorithm performance and quality metrics. Plos one, 17(4), e0266325. https://doi.org/10.1371/journal.pone.0266325

Saparuddin, S., Kusumaningrum, A., Setiorini, A., & Takidah, E. (2025). The Rise Of The Beauty Industry In Indonesia: Assessing Its Contribution To National Economic Growth. International Journal Of Economic Literature, 2(10), 1013-1029.

Syaputri, D. R., & Herawati, S. (2025). Influence Of Product Quality On Repurchase Intention Mediated By Customer Satisfaction On Sunscreen Products In Cimahi City. BIMA Journal (Business, Management, & Accounting Journal), 6(2), 1377-1386. https://doi.org/10.37638/bima.6.2.1377-1386

Vuković, D. B., Dekpo-Adza, S., & Matović, S. (2025). AI integration in financial services: a systematic review of trends and regulatory challenges. Humanities and Social Sciences Communications, 12(1), 562. https://doi.org/10.1057/s41599-025-04850-8

Wang, S., Zhang, Y., Xiao, Y., & Liang, Z. (2025). Artificial intelligence policy frameworks in China, the European Union and the United States: An analysis based on structure topic model. Technological Forecasting and Social Change, 212, 123971. https://doi.org/10.1016/j.techfore.2025.123971

Weston, S. J., Shryock, I., Light, R., & Fisher, P. A. (2023). Selecting the number and labels of topics in topic modeling: A tutorial. Advances in Methods and Practices in Psychological Science, 6(2), 25152459231160105. https://doi.org/10.1177/25152459231160105