Sarcasm Analysis and Mood Retention Using NLP Techniques

Soumen Mukherjee, Arup Bhattacharjee, Srijita Majumdar, Debabrata Datta, Arpan Deyasi, Anal Acharya

OAI: oai:igi-global.com:289952 • DOI: 10.4018/IJIRR.289952

Sarcasm detection in written texts is the Achilles’ heel of research areas in sentiment analysis, especially with the absence of the rightful verbal tone, facial expression or body gesture that leads to random misinterpretations. It is crucial in sectors of social media, advertisements and user feedbacks on services that require proper interpretation for service evaluation and improvisation of their products. The objective here thereby is to identify sarcasm within a given text by experimenting with the original predicted mood of the text and work on its transformation with the several variations in combination of the standard sarcastic elements present in the corresponding writing. Here standard NLP techniques are used for identification and interpretation. This involves detecting primary connotation of the given text (e.g. positive/neutral/negative), followed by detecting elements of sarcasm. Then, under the presence of the sarcasm indicator algorithm, the rightful interpretation of the previously detected mood is attempted.

Information Retrieval Library and Information Science Information Retrieval