Exploring The Mechanism And Predicting The Evolution Of Cyberbullying On Social Network Sites

Social network sites are frequent venues for cyberbullying,and due to the openness and immediacy of social network sites,cyberbullying based on social network sites has gradually evolved into a continuous and dynamic process.With the evolution of cyberbullying,the magnitude,extent and duration of its impact can cause more serious damage to the victim.In order to organically combine theoretical mechanisms with actual control operations,this research mainly combines the empirical model with machine learning prediction method.Based on understanding the evolution mechanism of cyberbullying and predicting the extent of the evolution of cyberbullying,it is convenient to recognize the cyberbullying event that require prevention.Therefore,this research can effectively implement theoretical guidance-based prevention and control strategies for it.In the empirical model,we applied the heuristic system model(HSM)and added time as a moderator to further explain the dynamic mechanism.We collected 5 real cyberbullying incidents in Sina Weibo,which is the largest social network site in China.The results show that source attractiveness has positive impact on the evolution of cyberbullying.Content objectivity and amount of information negatively affect the evolution of cyberbullying.With the evolution of time,the impact of amount of information and source attractiveness on the evolution of cyberbullying will be weakened,and the influence of objectivity on the evolution of cyberbullying will be further enhanced.Based on the empirical research,we used XGBoosts algorithm,support vector machine(SVM)algorithm and GBDT algorithm to construct the classification model.It also helps predict the extent of evolution of cyberbullying accurately.From the dynamic point of view,this research studies the dynamic mechanism of the evolution of cyberbullying.And it enriches the application of HSM theory in cyberbullying research,which shows that the dual process of HSM appears.At the same time,this research provides a basis for platform and personal protection against cyberbullying and weakening the risk of cyberbullying.