HateGuard: A Real Platform for Cyberbullying Detection
Year: 2026
Type: article
Abstract: Cyberbullying on social media represents a significant challenge to online safety, with potentially harmful consequences for individuals and communities. Although advances in machine learning have improved the automated detection of abusive content, much of the existing research remains focused on offline classification using benchmark datasets, with comparatively limited attention to operational deployment. This paper presents HateGuard, an operational framework for on-demand harmful-content screening and supervisory monitoring of authorized Instagram interactions. The framework integrates authorised data acquisition through the Instagram Graph API, automated preprocessing, a previously validated machine-learning detection engine, persistent data management, contextual result presentation, interactive visualisation, and automated email notification within a unified web-based architecture. Rather than introducing an additional classification algorithm, HateGuard investigates how an experimentally validated detection methodology can be transformed into an operational monitoring and decision-support platform. Functional end-to-end validation demonstrates the integration of authentication, Instagram interaction retrieval, machine-learning inference, result visualisation, and post-analysis alert generation within a modular deployment environment. HateGuard therefore provides a practical foundation for investigating the transition from offline cyberbullying detection research towards secure, extensible, and operational social media monitoring systems.
Source: Applied Sciences
DOI: 10.3390/app162211328
Institution: Universidad de Granada
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,896
OpenAlex ID: W7234259088
