Detection of Cyberbullying Tweets in Twitter Media Using Random Forest Classification
Gunasekar Thangarasu, Kesava Rao Alla
Abstract
Users of social media platforms have become more vulnerable to violent crimes, cyberbullying, and hate speech as a direct result of their increased use of online platforms, which has also contributed to this vulnerability. Another factor that has contributed to this vulnerability is the fact that users of online platforms have increased their usage of these platforms. It is of the utmost importance for the prevention of cyberbullying (CB) in smart cities to design a self-sufficient anti-cyberbullying engine that can recognise cyberbullying texts in social media posts. This is one of the most important steps that can be taken. Those who live in smart cities won’t have to stress about the possibility of getting into altercations with their neighbours because of this, which will allow them to unwind and take more joy in life. In this paper, a state-of-the-art automatic classification model that can recognise CB texts without requiring them to be constrained into a specific shape in high-dimensional space. This model was conceived of and created. Due of these constraints, we designed a text classification engine that first pre-processes the tweets by removing noise and other background information. Then, it classifies the data without overfitting the data by extracting the necessary features and then classifying the data. The random forest classifier can obtain a higher level of accurate classification than other types of classifiers, such as the conventional classifiers. The findings of the validation indicate that the random forest classifier, which has increased text classification accuracy, gives accurate classification results.