Distinguishing between factual information and insulting or abusive messages bearing words or phrases in news articles

Abstract

Since Internet has become the leading source of information for the users, flames or abusive messages have also become the prominent factors of time wasting for retrieving information. Moreover, a text can contain factual information as well as abusive or insulting contents. This paper describes a new approach for an automated system to distinguish between information and personal attack containing insulting or abusive messages in a given document. In NLP, flames or abusive messages are considered as extreme subjective language, which refers to detect personal opinions or emotions in a news article. Insulting or abusive messages are viewed as extreme subset of the subjective language because of its extreme nature. We defined some rules to extract the semantic information of a given sentence from the general semantic structure of that sentence.