The enormous use of sarcastic text in all forms of communication in social media will have a physiological effect on target users. Each user has a different approach to misusing and recognising sarcasm. Sarcasm detection is difficult even for users, and this will depend on many things such as perspective, context, special symbols. So, that will be a challenging task for machines to differentiate sarcastic sentences from non-sarcastic sentences. There are no exact rules based on which model will accurately detect sarcasm from many text corpus in the current situation. So, one needs to focus on optimistic and forthcoming approaches in the sarcasm detection domain. This paper discusses various sarcasm detection techniques and concludes with some approaches, related datasets with optimal features, and the researcher's challenges.

翻译：在社交媒体各类交流形式中，讽刺性文本的广泛使用会对目标用户产生心理影响。每位用户在滥用和识别讽刺时都有不同的方式。即使对用户而言，讽刺检测也十分困难，这取决于视角、语境、特殊符号等诸多因素。因此，对于机器而言，区分讽刺性句子与非讽刺性句子将是一项具有挑战性的任务。在现有情况下，没有任何确切的规则能确保模型从大量文本语料中准确检测出讽刺。因此，我们需要关注讽刺检测领域中乐观且可行的未来方法。本文讨论了多种讽刺检测技术，总结了一些方法、相关数据集及最优特征，并指出了研究人员面临的挑战。