,
Research Scholar, Department of ECM, Koneru Lakshmaiah Education Foundation , Vaddeswaram, Guntur, Andhra Pradesh , India
Professor & Coordinator (FED), Koneru Lakshmaiah Education Foundation , Vaddeswaram, Guntur, Andhra Pradesh , India
Sentiment analysis (SA) is important in comprehending the opinions and online discussions of people in different languages, particularly in high-resource and low-resource languages. In the present paper, a new model of transfer deep learning is presented, which incorporates cross-lingual embeddings in an attempt to boost sentiment analysis in both high and low-resource languages. HRLs get access to a vast number of linguistic resources and datasets, and LRLs have to grapple with the lack of labeled data, which makes the classic models of machine learning unproductive. The suggested framework uses transfer learning (TL) to mitigate these difficulties so that models trained on HRLs could be applied to LRLs, which would help in the problem of data scarcity. The model is a hybrid deep learning model, which uses both pre-trained models and cross-lingual embeddings, that can enhance the sentiment classification performance of both HRLs and LRLs. The comprehensive experiments of the multilingual data, as well as Twitter data, testify to the fact that the offered method is more effective than the traditional models. As an example, the proposed method has an accuracy of 96.42%, which is higher than Bi-LSTM and Bi-GRU with 91.45% and 88.57% accuracy, respectively. This performance depicts the efficiency of the suggested methodology in sentiment analysis in under-resourced languages. The findings show that the transfer learning can be used to substantially expand multilingual sentiment analysis through the use of transfer learning. These models can be further optimized to produce regionally specific languages in future studies and make them more scalable.
This is an open access article distributed under the Creative Commons Attribution Non-Commercial License (CC BY-NC) License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
0
The statements, opinions and data contained in the journal are solely those of the individual authors and contributors and not of the publisher and the editor(s). We stay neutral with regard to jurisdictional claims in published maps and institutional affiliations.