Real-Time Personalized Detection of Cyberbullying and Online Harassment in Social Networks Using Jetson Nano
Jackson Diaz-Gorrin, Pino Caballero‐Gil, Cándido Caballero‐Gil · 2024
This research focuses on analyzing social network comments to identify and mitigate potential cyberbullying. The primary objective is to develop a robust monitoring system that serves as a parental control tool, adding a crucial layer of protection to shield children from harmful and inappropriate messages in their online interactions. To achieve this, the study addresses a complex classification problem by evaluating and comparing various machine learning algorithms designed to categorize comments as either toxic or non-toxic. The most effective classifier is then implemented on a low-power device, specifically the NVIDIA Jetson Nano, to enable continuous and efficient monitoring of social media comments that can be customized and updated for every user. This deployment ensures that the system operates consistently without significant computational overhead, making it well-suited for real-world applications across all types of devices where ongoing vigilance is essential. Ultimately, the integration of this monitoring system aims to provide a practical solution for enhancing online safety and fostering a healthier digital environment for everyone, but especially for susceptible groups such as children.