Cyberbullying classification using text mining

Cyberbully is a misuse of technology advantage to bully a person. Cyberbully and its impact have occurred around the world and now the number of cases are increasing. Cyberbullying detection is very important because the online information is too large so it is not possible to be tracked by humans. The purpose of this research is to construct a classification model with optimal accuracy in identifying cyberbully conversation using Naive Bayes method and Support Vector Machine (SVM) then applying n-gram 1 to 5 for the number of class 2, 4, and 11 for each method. Naive Bayes yields an average accuracy of 92.81%, SVM with a poly kernel yields an average accuracy of 97.11% It can be concluded that SVM with poly kernel yields higher accuracy than SVM with other kernels, Naive Bayes, and Kelly Reynolds research method of decision tree (J48) and k-NN.