Self-attention bidirectional long Short-Term memory assisted natural language processing on sarcasm detection and classification in social media platforms.

Sarcasm is a form of irony that expresses negative opinions. Sarcasm poses a linguistic problem owing to its symbolic nature, where deliberate meaning challenges correct understanding. Sarcasm is more common on social media and in day-to-day life. Sarcasm detection in written text is a challenge that has attracted the attention of several researchers. Therefore, sarcasm is an essential process in natural language processing (NLP). This study discusses the ideas of sarcasm and its significance in present sarcasm research. The automated model of sarcasm recognition includes selecting appropriate approaches, selecting a dataset, and preprocessing steps involving Transformer architectures, a rule-based approach, deep learning (DL), and machine learning (ML) models. This manuscript proposes a Sarcasm Classification and Detection using NLP on Social Media Platforms (SDCNLP-SM) technique. The objective of the SDCNLP-SM technique is to effectively and automatically recognise sarcastic text. To accomplish this, the SDCNLP-SM technique performs data preprocessing and a Word2Vec-based word-embedding step. Finally, the Self-Attention with Bidirectional Long Short-Term Memory (SA-BLSTM) model is employed for sarcasm classification. The comparison analysis of the SDCNLP-SM model showed a superior accuracy of 94.45% compared to existing models on the headline dataset.