Classification of Online Pernicious Comments using Machine Learning

Internet has become the biggest platform to represent our skills. Various websites allow people to use their platform to showcase their skills through videos, articles and other information in different formats. Most of the websites provide facility of commenting on any of uploaded information. However, there is possibility that people can use abominable language in their comments. This paper mainly focuses on identification of these inadequate comments and classification of these into different categories. The required data is taken from machine learning site ‘Kaggle’ (www.kaggle.com). The comments are classified into 6 different categories- toxic, severe toxic, obscene, threat, insult and identity hate. We have used four different machine learning algorithms- logistic regression, support vector machine (SVM), K nearest neighbour and decision tree. All these mentioned models successfully predict classes of the comments. Out of these models support vector machine gives best result.