Paper Title
A Survey on Sarcasm Detection

Abstract
Sarcasm requires some common learning among speaker and audience; it is a significantly logical marvel. Most computational ways to deal with sarcasm identification, in any case, regard it as a simply etymological issue, utilizing data, for example, lexical signs and their relating assessment as prescient features. Sarcasm detection is a vital procedure to filter out boisterous information (for this situation, snide sentences) from preparing information inputs, which can be utilized for natural language sentence generation. This paper is an aggregation of past work in sarcasm detection. In this paper it is demonstrated that by including an additional semantic data from the setting of an articulation on Twitter– for example, properties of the author, the gathering of people and the prompt informative condition – can accomplish gains in exactness contrasted with absolutely phonetic highlights in the detection of this unpredictable marvel, while likewise revealing insight into highlights of relational association that empower sarcasm in discussion. Keywords - Machine Learning, Natural Language Processing (NLP), Sarcasm, and Twitter