GAN-BElectra: Enhanced Multi-class Sentiment Analysis with Limited Labeled Data

Performing sentiment analysis with high accuracy using machine-learning techniques requires a large quantity of training data. However, getting access to such a large quantity of labeled data for specific domains can be expensive and time-consuming. These warrant developing more efficient techniques that can perform sentiment analysis with high accuracy with a few labeled training data. In this paper, we aim to address this problem with our proposed novel sentiment analysis technique, named GAN-BElectra. With rigorous experiments, we demonstrate that GAN-BElectra outperforms its baseline technique in terms of multiclass sentiment analysis accuracy with a few labeled data while maintaining an architecture with reduced complexity compared to its predecessor.