Automating the Detection of Sarcastic Statements in Natural Language Text

摘要

This article is devoted to the sarcasm recognition in the text written in a natural language. The main goal is to increase the accuracy of sentiment analysis. The sentiment level determination of a text that describes the appearance of a person was chosen as a domain area for the experiment. At first, references to the personality and elements that describes appearance from text are detected using the method of latent semantic analysis. The next step is to evaluate the attitude to a person in text using pre-labeled sentiment dictionary. At this stage, the method of recognizing sarcastic sentences that contains a description of the appearance is used. The sentiment level should be re-evaluated in the person information model. The results of the experiment showed that the recognition of sarcasm based on the morphological features of words and the frequency characteristics of the sentences does not effectively increase the accuracy of sentiment level determination.