Automatic Detection and Blocking of Cyberbullying by Relatedness Calculation and Clustering
Priya S M, Dr. Asha T
Cyberbullying Detection, K-means, Leacock and Chodorow, Reconstruction, Social Media.
Cyberbullying is emerging as a serious problem affecting children, young adults and adolescents. This is because of increasing popularity and usage of social media by the teenagers. It will differ from traditional bullying usually occurs in face-to-face communication during school and college days. In this work, we Automatically detect and Block the cyber bullying messages to create a healthy and safe social media environment. We mainly work on creating an online social network, preparing a training corpus contains the bullying words and swear words, determine the Relatedness of the words using Leacock and Chodorow algorithm to detect the bullying posts. Our work is based on text representation learning and can exploit the hidden features or noisy data in the bullying information by Reconstruction.