Title: Sarcasm detection using enhanced glove and bi-LSTM model based on deep learning techniques

Authors: M. Anusha; R. Leelavathi

Addresses: PG & Research Department of Computer Science, National College (Autonomous), India; Affiliated to Bharathidasan University, Trichy, Tamil Nadu, India ' PG & Research Department of Computer Science, National College (Autonomous), India; Affiliated to Bharathidasan University, Trichy, Tamil Nadu, India

Abstract: The challenge of sarcasm detection in social media is heightened by its informal vernacular, the necessity for context, and the nuanced linguistic indicators that define sarcasm. This study explores the effectiveness of bi-LSTM models in conjunction with GLOVE and word2vec embeddings in identifying sarcastic expressions within social media platforms. This empirical evidence suggests that bi-LSTM models equipped with GLOVE embeddings surpass those with word2vec embeddings in all performance metrics, including accuracy, precision, recall, and F1-score. The bi-LSTM models demonstrate an improved ability to discern sarcasm's subtleties in text thanks to GLOVE's ability to encapsulate semantic meanings and contextual nuances. This research underscores the critical role of advanced pre-trained embeddings like GLOVE in elevating NLP tools for sarcasm detection on social media. Such advancements are particularly relevant to fields that depend on precise interpretation of user sentiment, such as brand monitoring, sentiment analysis, and broader social media analytics.

Keywords: sentiment analysis; sarcasm detection; natural language processing; NLP; word embedding; bi-LSTM; glove embeddings; deep learning techniques; enhanced glove.

DOI: 10.1504/IJIEI.2025.144268

International Journal of Intelligent Engineering Informatics, 2025 Vol.13 No.1, pp.26 - 54

Received: 15 Jun 2023
Accepted: 28 May 2024
Published online: 04 Feb 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article