Machine Learning Solutions for controlling Cyberbullying and Cyberstalking

55 ABSTRACT: In the last few years cybercrimes take different forms of shapes to propagate unwanted activities. Such email transactions are used by criminals as a means to initiate cybercrimes that inlcudes phishing, spamming, cyberbullying and cyberstalking. Cyberstalking is a relatively new surfacing cybercrime, which is now extensively used by criminals and antisocial elements. Combating email-based cyberstalking is a challenging task that includes the approaches such as -a robust method for filtering and detecting cyberstalking emails and documenting evidence for detecting such cybercrimes and to prevent such actions. We have initiated a research exercise which involves machine learning approach to find and control file evidence. Our mechanisms of research is based on a novel robust feature selection approach to select informative features, aiming to improve the performance.