Social Sentiment Sensor in Twitter for Predicting Cyber-Attacks Using ℓ₁ Regularization.

Area of Science:

• Data Science
• Cybersecurity
• Social Media Analysis

Background:

• Online social media serves as a key channel for communication and expression, generating vast amounts of data.
• Sentiment analysis, utilizing Natural Language Processing and Machine Learning, is employed to understand user opinions and predict real-world events from social media data.
• Cyber-attacks are increasingly linked to online social network dynamics, with hacker activists motivated by social events.

Purpose of the Study:

• To develop a methodology for tracking social data that can potentially trigger cyber-attacks.
• To enable the monthly prediction of tweets containing content related to security attacks.
• To detect security incidents based on analyzed social media data.

Main Methods:

• Leveraging data science techniques to analyze online social media information, specifically Twitter data.
• Employing sentiment analysis, Natural Language Processing (NLP), and Machine Learning (ML) for opinion interpretation.
• Utilizing L1 regularization for the prediction of security attack-related tweets and incident detection.

Main Results:

• A novel methodology for tracking social data indicative of cyber-attack triggers has been developed.
• Successful monthly prediction of tweets related to security attacks was achieved.
• Security incidents were detected through the analysis of social media content.

Conclusions:

• Social media data contains valuable indicators for predicting and understanding cyber-attack motivations and activities.
• The proposed methodology, incorporating L1 regularization, offers a reliable approach for proactive cybersecurity.
• This research highlights the intersection of social phenomena, online behavior, and cybersecurity threats.