| Enhancing the Grasshopper Optimization with Deep Learning Based on Sentiment Analysis on Social Media | |
|
MSRDG International Journal of Computer Scientific Technology & Electronics Engineering
© 2026 by MSRDG IJCSTEE Journal
Volume 2 Issue 2
Year of Publication: 2026 |
Paper Download Article ID MSRDG-IJCSTEE-V2I2P105 |
|
Abstract: Sarcasm detection in social media text is a fundamentally challenging natural language processing (NLP) task owing to the inherently ambiguous and context-sensitive nature of ironic expressions. Conventional sentiment analysis systems frequently misclassify sarcastic statements because the surface-level polarity of words contradicts their intended meaning. This paper presents a novel hybrid framework that integrates the Grasshopper Optimization Algorithm (GOA) with a deep learning model combining Bidirectional Long Short-Term Memory (Bi-LSTM) networks and a Convolutional Neural Network (CNN) augmented with an attention mechanism for effective sarcasm detection on social media platforms. The GOA is employed to optimize the hyperparameters of the deep learning model, thereby mitigating the problem of manual parameter tuning and improving convergence efficiency. Experimental evaluations are conducted on two benchmark datasets, namely the MUStARD dataset and the iSarcasm Twitter corpus, demonstrating that the proposed GOA-BiLSTM-CNN model achieves superior performance with an accuracy of 93.7%, precision of 92.8%, recall of 93.1%, and F1-score of 92.9%, outperforming several state-of-the-art approaches including BERT-base, standalone Bi-LSTM, and SVM classifiers. The proposed framework contributes a computationally efficient and scalable solution for fine-grained sarcasm-aware sentiment analysis that is well-suited for real-world deployment in opinion mining, brand monitoring, and social listening applications. |
|
| Keywords: Sarcasm Detection, Grasshopper Optimization Algorithm, Deep Learning, Bi-LSTM, CNN, Sentiment Analysis, Social Media, Natural Language Processing | |
