Education, Science, Technology, Innovation and Life
Open Access
Sign In

Design and Implementation of a Sentiment Analysis System Based on Deep Learning

Download as PDF

DOI: 10.23977/autml.2026.070113 | Downloads: 4 | Views: 137

Author(s)

Chuwei Wang 1, Zicheng Wang 1, Xihe Wang 1, Mingxing Wang 1

Affiliation(s)

1 School of Electronic and Information Engineering, University of Science and Technology Liaoning, Anshan, China

Corresponding Author

Mingxing Wang

ABSTRACT

Sentiment analysis, a core NLP task, has important application value in public opinion monitoring, product evaluation, and social media analysis. Traditional methods based on manual feature extraction or shallow learning models have limitations in complex text scenarios. To address these, this paper designs and implements a sentiment analysis system based on an improved RoBERTa model. The system selects three real public datasets, optimizes data quality through text cleaning, normalization, and back - translation augmentation. It integrates an emotional attention mechanism and introduces FGSM adversarial training. Ablation and comparative experiments are conducted. Results show the proposed model achieves an average F1 - score of 94.2% on the three datasets, 3.7 to 5.1 percentage points higher than the baseline model. The system has strong generalization, high robustness, and reliable performance, offering a practical solution for multi - scenario text sentiment analysis. 

KEYWORDS

Sentiment Analysis, RoBERTa Model, Adversarial Training, Ablation Experiment, Pre-trained Language Model, Text Augmentation, Emotional Attention Mechanism

CITE THIS PAPER

Chuwei Wang, Zicheng Wang, Xihe Wang, Mingxing Wang. Design and Implementation of a Sentiment Analysis System Based on Deep Learning. Automation and Machine Learning (2026). Vol. 7, No. 1, 104-111. DOI: http://dx.doi.org/10.23977/autml.2026.070113.

REFERENCES

[1] Prova I N N ,Ravi V ,Singh P M , et al. Multilingual sentiment analysis in e-commerce customer reviews using GPT and deep learning-based weighted-ensemble model[J].International Journal of Cognitive Computing in Engineering,2026,268.DOI:10.1016/J.IJCCE.2025.10.003.
[2] Tigani S . Next-gen HR interview process parametric AI copilot design based on pretrained LLM and sentiment analysis deep learning models[J].Pattern Analysis and Applications,2025,28(4):205.DOI: 10.1007/S10044-025-01581-Z.
[3] Tarik N ,Jadhav A . DStaL: Multilevel Fusion Classifier Based Analysis of Multimodal Sentiments Using Deep Learning Models[J].Computational Intelligence,2025,41(6):e70154.DOI:10.1111/COIN.70154.
[4] Yarjanli M ,Mahdinasab N . A Hybrid Deep Learning Model for E‐Commerce Recommendations: Sentiment Analysis With Autoencoders and Generative Adversarial Networks[J].International Journal of Intelligent Systems,2025, 2025(1):3852068.DOI:10.1155/INT/3852068.
[5] Sitoula S ,Shahi B T ,Wibowo S , et al. Sentiment analysis of Nepali social media text with a hybrid deep learning model[J].Social Network Analysis and Mining,2025,15(1):85.DOI:10.1007/S13278-025-01508-W.
[6] Li L ,Choo C W ,Leong C Y , et al. Predicting purchase intentions in online food delivery using deep learning and AIDA model: Insights from sentiment analysis of user reviews[J].International Journal of Engineering Business Management,2025,17.DOI:10.1177/18479790251375827.
[7] Sharma R ,Kukreja V . Deciphering emotions in comics: analysis of emotion classification of comic characters via attention-based deep learning models[J].Multimedia Tools and Applications,2025,84(39):1-33.DOI:10.1007/S11042-025-21017-4.
[8] Talukder A M ,Uddin A M ,Roy S , et al. A hybrid deep learning model for sentiment analysis of COVID-19 tweets with class balancing.[J].Scientific reports,2025,15(1):27788.DOI:10.1038/S41598-025-97778-7.
[9] Yanping L . Building and Optimizing Deep Learning Models for Sentiment Analysis in English Text[J].Journal of Cases on Information Technology (JCIT),2025,27(1):1-17.DOI:10.4018/JCIT.382380.
[10] Atlas G L ,Arockiam D , Muthusamy A , et al.  A modernized approach to sentiment analysis of product reviews using BiGRU and RNN based LSTM deep learning models.[J].Scientific reports,2025,15 (1) :16642. DOI :10.1038/S41598-025-01104-0.

Downloads: 5198
Visits: 260985

Sponsors, Associates, and Links


All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.