A Language Model-Driven Semi-Supervised Ensemble Framework for Illicit Market Detection Across Deep/Dark Web and Social Platforms
Navid Yazdanjue, Morteza Rakhshaninejad, Hossein Yazdanjouei, Mohammad Sadegh Khorshidi, Mikko S. Niemelä, Fang Chen, Amir H. Gandomi · 2025
Read A Language Model-Driven Semi-Supervised Ensemble Framework for Illicit Market Detection Across Deep/Dark Web and Social Platforms by Navid Yazdanjue, Morteza Rakhshaninejad, Hossein Yazdanjouei, Mohammad Sadegh Khorshidi, Mikko S. Niemelä, Fang Chen, Amir H. Gandomi (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-22912.txt.
Description
Navid Yazdanjue and co-authors built a 2025 system to spot illegal markets hiding on the hidden web and social apps. Sellers of drugs, weapons, and stolen logins now work through dark sites, chat channels, and paste pages. Finding them is hard because labeled data is scarce and slang changes fast. The team's system sorts posts in two stages, first flagging sales posts, then sorting them by what is sold.
A modern language model reads each post for meaning, while hand-built features catch clues like coin addresses and emails. Three simpler models vote together in the first stage, with rare labels given extra weight. Across three data sets, the system beat six rival models on accuracy. The authors frame it as a tool for teams that police online crime.
Catalog record
| CREATOR | Navid Yazdanjue, Morteza Rakhshaninejad, Hossein Yazdanjouei, Mohammad Sadegh Khorshidi, Mikko S. Niemelä, Fang Chen, Amir H. Gandomi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-22912