VASSL

VASSL detects and supports labeling of social spambots on social media platforms by integrating visual analytics with dimensionality reduction, sentiment analysis, and topic modeling to aid detection and analysis.

Key Features:

• Visual analytics with connected views: Multiple connected views enable exploration and comparison of social media accounts and the selection and analysis of account groups.
• Dimensionality reduction: Employs dimensionality reduction techniques to manage and simplify large datasets and reveal patterns and anomalies.
• Sentiment analysis and topic modeling: Incorporates sentiment analysis and topic modeling to analyze content and context of posts for distinguishing automated behavior from human behavior.
• Scalability and performance enhancement: Designed to improve scalability and performance of manual labeling processes across large datasets.

Scientific Applications:

• Social media analytics and cybersecurity: Supports research on detection and characterization of malicious accounts in social media ecosystems.
• Behavioral analysis of spambots: Enables study of behavioral patterns and interaction profiles of spambots.
• Detection algorithm development: Provides analytical outputs that inform development and evaluation of spambot detection algorithms.
• Mitigation research: Aids investigation into strategies to mitigate the impact of spambots on digital communication platforms.

Methodology:

Combines visual analytics with multiple connected views, dimensionality reduction, sentiment analysis, and topic modeling.