Abstract :

The Rapid Growth Of Social Media And Digital Communication Platforms Has Significantly Increased The Volume Of Cybersecurity-related Information Shared In Real Time. Platforms Such As Twitter And Online Forums Generate Massive Amounts Of Unstructured Textual Data Containing Valuable Insights Into Cyber Threats, Vulnerabilities, And Ongoing Attacks. To Overcome These Limitations, This Study Proposes A Hybrid NLP-based Cyber Intelligence Framework For Real-time Threat Detection From Social Media Data. The Framework Begins With Preprocessing Steps, Including Tokenization, Normalization, And Noise Removal, To Enhance Data Quality. It Then Utilizes Sentence-Bidirectional Encoder Representations From Transformers (SBERT) To Generate Semantic Embeddings, Enabling A Deeper Understanding Of Contextual Relationships Within Textual Content. To Address Class Imbalance, The Synthetic Minority Over-sampling Technique (SMOTE) Is Applied, Ensuring Balanced Datasets For Effective Training. The System Integrates Multiple Classifiers, Including Stochastic Gradient Descent (SGD), Complement Naïve Bayes (CNB), And A Hybrid Model Combining Dense Neural Networks (DNN) With Linear Discriminant Analysis (LDA). This Multi-model Architecture Improves Feature Representation And Classification Performance. Experimental Results Demonstrate Enhanced Accuracy And Reliability, Offering A Scalable And Efficient Solution For Automated Cyber Threat Intelligence Using Large-scale Social Media Data.

Published:

09-4-2026

Issue:

Vol. 26 No. 4 (2026)

Page Nos:

2057-2066

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite