Abstract:
"The widespread use of digital communication has made it more difficult to discern subtleties in
text, particularly sarcasm, which mostly depends on tone and context that are missing from digital
interactions. In order to tackle the complex issue of sarcasm recognition and response creation,
this thesis introduces ""SarcaBot,"" a novel hybrid approach that combines Natural Language
Processing (NLP) and Text Transformer models. Emojis are added to the system to enhance the
response, which mimics human-like replies in online conversations. Because sarcasm depends on
the speaker's meaning and context, its linguistic nuance presents substantial hurdles for automated
systems and frequently leads to misinterpretations that impede efficient digital communication.
While Text Transformers offer potential contextual comprehension, their application in sarcasm
detection is still underexplored. Traditional NLP models are unable to capture these subtleties. In
order to close this gap and produce responses that are appropriate for the context, this research
combines the best features of Text Transformers with NLP techniques to increase the accuracy of
sarcasm recognition."
