Abstract: More than 1.96 billion are bound to have an inevitable social life. However, the growing decade poses serious challenges and the online-behaviour of users have been put to question. Increasing cases of harassment and bullying along with cases of fatality have been a serious issue. Though, many old-school models are available to control the mishap, the need to effectively classify the bullying is still feeble. To effectively monitor the bullying in the virtual space and to stop the deadly aftermath with implementation of Machine Learning and Language processing. In this paper, we propose a methodology to provide a binary classification of cyberbullying. Our method uses an innovative concept of CNN for text analysis however the existing methods use a naive approach to provide the solution with less accuracy. An existing twitter dataset is used for experimentation and our framework is verified with other existing procedures and is found to provide better accuracy and classification.
Keywords: Convolution Neural Network (CNN), Cyberbullying, Latent Semantic Analysis (LSA) and Latent Dirichlet Allocation (LDA)
|
DOI:
10.17148/IJARCCE.2018.71215
