Detecting Cyber Bullying using Ml Algorithm

Cyberbullying is a crime in which a victim attacks a person with hatred and online threats. In order to recognize word similarity in bullies' tweets and allow use of machine learning, an ML model can be built to automatically identify social network bullying behavior. Nonetheless, several social media abuse detection approaches have been adopted, but many of them are textually focused. To recognize and avoid bullying on Twitter, a machine learning algorithm is recommended. 4 classifiers, i.e. SGD, Naïve Bayes, Random Forest and Logistic Regression are used for the preparation and testing of social media abuse content. All four Logistic Regression, Random Forest, Naive Bayes and SGD (Stochastic Gradient Descent) were able to detect the true positives with 77 percent, 93 percent, 75 percent and 79 percent accuracy respectively.