Title: Metaheuristic-assisted deep ensemble technique for identifying sarcasm from social media data

Authors: Geeta Abakash Sahu; Manoj Hudnurkar

Addresses: Faculty of Computer Studies (FoCS), Symbiosis International University (SIU), Symbiosis Centre for Research & Innovation (SCRI), Pune, India ' Operations and IT, Symbiosis International University (SIU), Symbiosis Centre for Management & Human Resource Development (SCMHRD), Pune, India

Abstract: Sarcasm is regarded as the enveloping linguistic factor in online documents that describes the deeply-felt subjective and opinions. This paper intends to introduce a sarcasm detection model that classifies words under sarcastic or non-sarcastic forms. Pre-processing is the initial phase, where the stop word removal and tokenisation are performed. The pre-processed data is then subjected to extracting the features, where, 'information gain, chi-square, mutual information, and symmetrical uncertainty-based features' are extracted. As the curse of dimensionality becomes the greatest crisis, optimal feature selection is carried out. For sarcasm detection, an ensemble classifier such as NN, RF, SVM and optimised DCNN is used, in which the weights of DCNN are optimally selected. For optimal feature selection and optimised DCNN, a hybrid optimisation model termed as Clan Updated Grey Wolf Optimisation (CU-GWO) is proposed. Finally, the effectiveness of the proposed algorithm is compared with extant methods in terms of various measures.

Keywords: sarcasm; tokenisation; information gain; optimised DCNN; CU-GWO optimisation.

DOI: 10.1504/IJWMC.2024.136558

International Journal of Wireless and Mobile Computing, 2024 Vol.26 No.1, pp.25 - 38

Received: 19 Apr 2022
Received in revised form: 22 Feb 2023
Accepted: 22 Feb 2023
Published online: 07 Feb 2024 *

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