Emerging Trends and Key Themes in Deep Learning Sentiment Analysis + A Bibliometric Analysis from 2014 to 2025
Keywords:
Deep Learning, Sentiment Analysis, Bibliometric Analysis , NLPAbstract
This study examines deep learning (DL) applications in sentiment analysis using Natural Language Processing (NLP) through a bibliometric analysis of 222 publications from Scopus databases, spanning 2024 to 2025. Key themes such as "sentiment analysis," "deep learning," and "Natural Language Processing" are analyzed, highlighting advancements across domains like healthcare, social media, and finance.
The research synthesizes progress in text classification and sentiment analysis, focusing on underrepresented languages and specific contexts. Techniques such as TCAODL-ANA, combining attention-based bidirectional gated recurrent units (ABiGRU) with the Aquila optimizer, excel in Arabic news article classification. Models like Arb-MCNN-Bi and reinforcement learning (RL) approaches enhance scalability and accuracy, preserving contextual nuances. Social media sentiment analysis benefits from ensemble methods and conditional GANs, while hybrid frameworks improve fake news detection and customer feedback assessment. Financial lexicons like XLex integrate transformer-based learning to balance interpretability and efficiency.
These findings underscore the integration of machine learning, deep learning, and hybrid methodologies in advancing NLP applications, offering significant insights for future research in e-commerce, finance, healthcare, and social discourse analysis within this dynamic field.
Downloads
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Sociolytics Journal
This work is licensed under a Creative Commons Attribution 4.0 International License.
The copyright in the text of individual articles (including research articles, opinion articles, book reviews, conference proceedings and abstracts) is the work of the authors, subject to a general license granted to SJ by transferring the copyright upon submission of articles and a Creative Commons CC-BY license granted to all others, as specified below. All work submitted to SJ implies that it presents original, unpublished work of the authors, which has not been published previously, has not been accepted for publication elsewhere and is not under consideration for publication elsewhere. Upon acceptance of the manuscript for publication, the Corresponding Author, on behalf of all authors of the manuscript, will be required to sign and submit a completed Copyright Form. As an author or contributor you grant permission to others to reproduce your articles, including any graphics and third-party materials supplied by you, under the Journal Terms and Conditions and subject to any copyright notices which you include in connection with such materials. A Creative Common Attribution (“CC BY”) license is granted to third parties. The current version is CC-BY, version 4.0 (http://creativecommons.org/licenses/by/4.0/), and the license will automatically be updated as and when updated by the Creative Commons organization. The following summaries are on the Creative Commons website as of 21st August 2016. The Creative Commons Attribution License 4.0 provides the following summary (where ‘you’ equals ‘the user’): Sharing — copying and redistributing the material in any medium or format and adapting — remixing, transforming, and building upon the material for any purpose, even commercially. The licensor cannot revoke these freedoms as long as there is full adherence to the license terms under the following terms: Attribution — you must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner but not in any way that suggests that the licensor endorses you or your use. No additional restrictions — you may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.
