A TRANSPARENT MACHINE LEARNING FRAMEWORK FOR SPAMBOT AND FAKE FOLLOWER DETECTION

Abstract

The rapid growth of social networking platforms has led to an increasing presence of spambots and fake followers, which significantly undermine the credibility, security, and user experience of online ecosystems. These malicious entities are often used to spread misinformation, manipulate public opinion, and artificially inflate popularity metrics, posing serious challenges for both platform providers and users. Traditional detection mechanisms rely on rule-based or black-box machine learning models that lack transparency and adaptability, making them insufficient to handle evolving bot behaviors. This paper proposes an interpretable AI-based machine learning framework for the identification of spambots and fake followers on social networks. The proposed approach integrates feature engineering, behavioral analysis, and interpretable models such as decision trees, SHAP (SHapley Additive exPlanations), and LIME (Local Interpretable Model-Agnostic Explanations) to ensure both high detection accuracy and model transparency. The system analyzes various attributes including user activity patterns, follower–following ratios, posting frequency, content similarity, and engagement metrics to distinguish between genuine users and malicious accounts. Experimental results demonstrate that the proposed model achieves superior performance in terms of accuracy, precision, recall, and F1-score compared to traditional methods while also providing clear explanations for its predictions. The interpretability aspect enhances trust and allows administrators to understand the reasoning behind classification decisions, making the system more reliable and actionable. Furthermore, the framework is scalable and adaptable to different social media platforms. This research contributes to the development of transparent and effective AI-driven solutions for combating social media manipulation and improving digital trust.