Talk · International Conference on Cybersecurity, Digital Forensics, and AI Applications · June 1, 2026
Cyber Threat Risk Assessment from Hacker Forum Emotion and Activity Signals
Сакен Мамбетов, Askar Khikmetov, Serik Joldasbayev, Bakhytzhan Kulambayev, Kuanysh Dossanbek, Tolganay Chinibayeva
Abstract
The rapid growth of cyber threats requires new approaches for the early detection of dangerous activity in online environments. Traditional cyber threat intelligence systems mainly rely on technical indicators such as vulnerability databases, intrusion alerts, and incident reports. However, hacker forums often contain early signals of vulnerabilities, exploit discussions, and emerging malicious activity before such information appears in formal security sources. This study proposes a cyber threat assessment approach based on hacker forum discussions. The method combines text preprocessing, language identification, rule-based emotion labeling, forum activity analysis, and temporal smoothing to construct two aggregated indicators: the Threat Emotion Index (TEI) and the Cyber Threat Risk Index (CTRI). The dataset consists of 589,788 forum posts collected from six sources. Language analysis showed that 534,197 posts were in English and 41,090 in Russian, confirming the multilingual nature of the observed discourse. Emotion signals were identified using six categories: anger, fear, joy, sadness, sarcasm, and neutral. TEI was computed as a weighted combination of emotional categories, while CTRI integrated emotional intensity with normalized forum activity. EWMA smoothing was applied to the CTRI time series to reveal sustained periods of elevated risk. The results show that the proposed framework provides an interpretable baseline for emotion-aware cyber threat intelligence and helps translate temporal changes in hacker forum discussions into aggregated threat indicators. Combining activity patterns, emotional signals, and temporal analysis can improve the early detection of cyber threat dynamics in multilingual online communities.
More by Serik Joldasbayev
- COMPARISON OF YOLOV8N AND YOLOV12N MODELS FOR SCALABLE REAL TIME VIDEO MONITORING OF CONSTRUCTION SITESScientific Journal of Astana IT University · 2026
- Design and Empirical Evaluation of a Four-Layer AI Agent Architecture for Automated Web Application Security TestingEngineering Technology & Applied Science Research · 2026 · Citations: 2
- AUTOMATION OF THE SYSTEM FOR INDUSTRIAL PRACTICE AND INTERNSHIPS FOR STUDENTS IN ORGANIZATIONS OUTSIDE OF THE UNIVERSITYNEWS OF THE NATIONAL ACADEMY OF SCIENCES OF THE REPUBLIC OF KAZAKHSTAN · 2025 · Citations: 1
- Graph-based neural networks with neural ODEs for robust speech processing in environmental and human-centric systemsE3S Web of Conferences · 2025 · Citations: 1
