Abstract

Natural Language Processing (NLP) is a broad field of computer science which attempts to automatically extract information from natural language texts.

This talk will focus on the problem of detecting the use of sarcasm in social media. Sarcastic sentences are constructed to initially appear to have a particular meaning, but the underlying intent of the message is very different. This presents a challenge to traditional sentiment analysis, which may fail to detect the true sentiment.

In this talk, I will introduce some general ideas in NLP , such as the concept of Part-of-speech (POS) tagging. I will then introduce the problem of sarcasm detection, and present the method and results of two papers addressing the issue. The first paper attempts to build a system which detects a type of sentence structure hypothesised to strongly indicate sarcasm. The second paper evaluates the usefulness of particular sets of POS tags in determining sarcastic utterances. I will finally compare the approaches of both papers.