Dark web data classification and threat response methods through topic modeling evaluation
Journal of Data Forensics Research · 30.06.2026 · 3 (1) · 27-54
DOI: 10.12972/jdfr.2026.3.1.2 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Data Forensics Research, online erschienen am 30.06.2026. Verfasst von Soonwook Kim und 6 weiteren Autor:innen.
Abstract (Originalsprache)
The dark web is an anonymous network in which criminal-related information, including hacking tools, financial fraud data, and illicit content, is actively distributed. Owing to its large-scale and unstructured nature, effective classification and analysis remain challenging. This study compared and evaluated topic-modeling techniques to identify threat categories and support threat-response strategies. A dataset was constructed from dark web HTML data collected over eight months. The data were preprocessed, including duplicate removal, content extraction, language filtering, normalization, and stop-word removal. Subsequently, topic classification was performed using Bag-of-Words (BoW)-based models (latent dirichlet allocation and non-negative matrix factorization), embedding-based models (BERTopic and Top2Vec), and the large language model-based topic generative pre-trained transformer framework. The analysis identified major categories, such as seized or closed sites, financial fraud, cryptocurrency-related activities, marketplaces, hacking services, and pornographic content. Embedding-based models distinguished more fine-grained categories through semantic understanding, whereas BoW-based models provided higher interpretability. These findings demonstrate the potential of topic modeling for large-scale dark web-data classification and cyber-threat intelligence applications.
Zitieren
APA
Kim, S., Oh, S., Choi, M., Kim, H., Yang, J., Kang, W., & Kim, G. (2026). Dark web data classification and threat response methods through topic modeling evaluation. Journal of Data Forensics Research, 3(1), 27-54. https://doi.org/10.12972/jdfr.2026.3.1.2
Vancouver
Kim S, Oh S, Choi M, Kim H, Yang J, Kang W, et al. Dark web data classification and threat response methods through topic modeling evaluation. Journal of Data Forensics Research. 2026;3(1):27-54. doi:10.12972/jdfr.2026.3.1.2
BibTeX
@article{Kim2026,
author = {Kim, Soonwook and Oh, Sojung and Choi, Minyoung and Kim, Hyungbin and Yang, Jueun and Kang, Woohyun and Kim, Gibum},
title = {Dark web data classification and threat response methods through topic modeling evaluation},
journal = {Journal of Data Forensics Research},
year = {2026},
volume = {3},
number = {1},
pages = {27--54},
doi = {10.12972/jdfr.2026.3.1.2}
}
EUVIMED