Sarcasm detection in social media using deep contextual models
1 Department of Computer Applications (MCA)Davangere University, Davangere
2 Department of central library, Davangere University,Davangere
3 Department of studies in Computer Science Davangere University, Karnataka, India.
Kumar Siddamallappa U; Orcid :-0000-0002-1978-3868
Neelamma G; Orcid :-0000-0003-1824-4858
Sowmya P; Orcid :-0009-0007-9059-9262
Research Article
World Journal of Engineering and Technology Research, 2026, 04(01), 001–007
Article DOI: 10.53346/wjetr.2026.4.1.0023
Publication history:
Received on 14 June 2026; revised on 20 July 2026; accepted on 22 July 2026
Abstract:
Social media platforms are filled with millions of posts created by users that express opinions, emotions, and experiences. Many of these posts have sarcasm in them where the intended meaning is different from the literal meaning, which makes sentiment analysis difficult. Conventional machine learning approaches often fail to capture the contextual and semantic information required for accurate sarcasm detection. Here we propose a deep contextual model based on RoBERTa and LSTM to identify sarcastic and non-sarcastic comments from social media text. The system performs text preprocessing, contextual embedding generation and classification to enhance the accuracy of sarcasm detection. Experimental results show that deep contextual models can outperform traditional machine learning methods and improve sentiment analysis by effective identification of hidden emotions
Keywords:
Sarcasm Detection; Sentiment Analysis; Social Media Analysis; Hidden Emotion Recognition
Full text article in PDF:
Copyright information:
Copyright © 2026 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0