Fake User Identity Detection Among Human and Bot Using Machine Learning Approach of Support Vector Machine

The rapid increase in fake identities and automated bots in online streams, especially on social media and e-commerce platforms, poses significant security and trust challenges. This paper presents a Support Vector Machine (SVM)-based Fake User Identity Detection Model that differentiates between human users and bots by analyzing account details, activity logs, and behavioral data. Initially, the model incorporates a preprocessing pipeline to clean, normalize, and structure input data, ensuring optimal feature representation. Then, the Mutual Information (MI) is employed for feature selection, and categorizing the most relevant attributes using the SVM classifier. The model is deployed through a web-based user interface, where the frontend is built using HTML and CSS, and the backend is implemented in Python to handle data processing, model inference, and API integration. The system can seamlessly integrate into online platforms, providing real-time fake identity detection to enhance security and user trust. Trained on a publicly available Twitter bot detection dataset, the SVM model achieves a high classification accuracy of 99.85%, significantly improving detection performance over baseline methods.