Paper Contents
Abstract
Identifying sarcasm in written communication especially from social media can be complex due to opposite nature of sarcastic utterance. The existing methods for sentiment analysis are not very successful when it comes to the sarcastic tweets because the explicit and implicit connotations are not taken into consideration. This work specifically aims at employing the Support Vector Machine (SVM) algorithm for enhanced classification of sarcastic and non-sarcastic text. Besides, the work covers text data preparation, feature extraction using TF-IDF, and word embeddings, and the training of an SVM model to achieve exact sarcasm classification. The study also looks at some of the advantages of SVM especially in conditions whereby data has higher dimensionality and in the generation of good margins. However, some of the problems that can be considered here are scalability of the model, the computational complexity, as well as the aspect of overfitting. Future work will improve the above-stated problems by finding efficient algorithms, using right hyperparameters and extracting efficient features. Further, the use of blockchain approach for maintaining the data authenticity and scalability using cloud computing for huge data set is suggested. These changes are envisaged to enhance the efficiency of the model for use in real-life applications such as monitoring the social media and sentiment analysis.
Copyright
Copyright © 2024 Geetika Arjel. This is an open access article distributed under the Creative Commons Attribution License.
