Towards Developing Effective Machine Learning Frameworks to Identify Toxic Conversations Over Social Media (original) (raw)

2018

Abstract

The advent of social media has great impact in the society and huge number of conversations are being made over social media everyday. Unfortunately, many of these conversations are meant for personal attacks or include abusive comments, which may have adverse effect to particular communities or individuals. The same may raise doubts about the liability and popularity of the social media forums, which should be prevented. A set of machine learning classifiers have been explored here to automatically identify abusive or toxic comments from social media posts. The empirical analysis on a data set of Wikipedia Talk Page using different such classifiers including a recurrent neural network has shown significant improvement towards

Tanmay Basu hasn't uploaded this paper.

Let Tanmay know you want this paper to be uploaded.

Ask for this paper to be uploaded.