Cyberbullying Detection in Schools Using Machine Learning: An Experimental Study on Instagram
Cyberbullying Detection in Schools Using Machine Learning: An Experimental Study on Instagram

Summary/Abstract: Cyberbullying is a growing problem in digital environments that affects students’ mental health. Anonymity, the broad reach of these platforms, and the lack of regulation worsen the situation. This research proposes that the use of machine learning (ML) can improve the detection of cases by aiming to increase the number of identified cases, reduce detection time, and enhance the accuracy rate. An applied and experimental methodology was implemented, comparing a control group (manual detection) with an experimental group (ML-based system). The results showed significant improvements across all indicators: The number of detected cases increased by 42.12%, detection time was reduced by over 99.9%, and the accuracy rate improved from 85.3% to 98.8%. These findings validate that the use of ML enhances the detection of cyberbullying cases, offering a scalable solution for educational institutions to transition from reactive to preventive strategies, thereby fostering safer digital ecosystems for students.

• Details
• Contents