Learning Deep on Cyberbullying is Always Better Than Brute Force

In this paper we present our research on detection of cyberbullying (CB), which stands for humiliating other people through the Internet. CB has become recognized as a social problem, and its mostly juvenile victims usually fall into depression, selfmutilate, or even commit suicide. To deal with the problem, school personnel performs Internet Patrol (IP) by reading through the available Web contents to spot harmful entries. It is crucial to help IP members detect malicious contents more efficiently. A number of research has tackled the problem during recent years. However, due to complexity of language used in cyberbullying, the results has remained only mildly satisfying. We propose a novel method to automatic cyberbullying detection based on Convolutional Neural Networks and increased Feature Density. The experiments performed on actual cyberbullying data showed a major advantage of our approach to all previous methods, including the best performing method so far based on BruteForce Search algorithm.