A Prediction Model for the Relationship Between Academic Stress and Cyberbullying Based on Ensemble Random Forest

With the widespread popularity of the Internet, cyberbullying among adolescents has become increasingly serious and has produced significant impact on their mental health. Academic stress and depressive mood are often thought to contribute to cyberbullying, but the exact relationship between these factors and cyberbullying has not been thoroughly studied. This study aims to clarify this relationship and use a machine learning algorithm called Ensemble Random Forest (ERF) to build a model that can predict cyberbullying behavior. We hope to provide schools and families with a scientific basis for intervention by using ERF to build a model that can predict cyberbullying behavior. For this study, we used machine learning methods focusing on the ERF algorithm to train the prediction model. Data came from a questionnaire survey of 300 secondary school students regarding their academic stress levels, depression levels, frequency of social media usage, and behavioral manifestations related to cyberbullying. We preprocessed and normalized the data and then evaluated the model's performance using cross-validation. The final results show that this model's predictive ability remains strong: the accuracy rate was 86.67%, the recall rate was 64.71%, the precision rate was 84.62%, and the F1 score was 73.33%.