Early detection of cyber-bullying from twitter data

Abstract

The vast volume of user created data in social media has created a wide-open path to data analytics in terms of extraction of features, analysis of polarity of information and classification of human emotional state. Sentiment analysis and data mining has reached at its peak in terms of its use in feature extraction and commercial analysis of products. Enormous approaches have been also seen in human behaviour analysis using ML (Machine Learning) algorithms and NLP (Natural Language Processing). To be very specific, it is not too hard to analyse data in granular level such as aspects. Social media play a distinct role in providing large dataset at no cost. And aspect-based sentiment analysis provides us the text analysis technique that categorisesuser provided data and assigns positive or negative value for the sentiment identified in it. In our approach we have tried to find out a solution of detection of cyber bullying in human interaction. This problem, we find a new threat in this digital era as commenting and replying to that comment play an important role in virtual communication, and just in this point, it takes no cost or hesitation factor to attack another person verbally or virtually.