A comparative study of e-commerce review sentiment analysis models based on VADER and RoBERTa

Yongli Bao

doi:10.54097/f6hyft52

Authors

Yongli Bao

DOI:

https://doi.org/10.54097/f6hyft52

Keywords:

Sentiment Analysis, VADER, RoBERTa, NLP

Abstract

The study compares the performance of two sentiment analysis models - VADER and RoBERTa - in analyzing Amazon product reviews. Using a dataset of reviews containing different star ratings, the study performed sentiment analysis on reviews using the VADER and RoBERTa models, respectively, and analyzed the performance of both on positive, neutral, and negative sentiment scores. The results show that the RoBERTa model performs well in capturing complex sentiment and contextual information, and especially has stronger recognition ability on extreme sentiments (e.g., 1-star and 5-star reviews.) The VADER model, although lightweight and fast, predicts weak correlation between sentiment scores and actual user ratings when dealing with complex semantics and long texts. This study provides reference value for sentiment analysis of user reviews on e-commerce platforms.

References

[1] Taboada, M. (2016). Sentiment analysis: An overview from linguistics. Annual Review of Linguistics, 2(1), 325-347.

[2] Baid, P., Gupta, A., & Chaplot, N. (2017). Sentiment analysis of movie reviews using machine learning techniques. International Journal of Computer Applications, 179(7), 45-49.

[3] Märkle-Huß, J., Feuerriegel, S., & Prendinger, H. (2017). Improving sentiment analysis with document-level semantic relationships from rhetoric discourse structures.

[4] Kamal, A. (2015). Review mining for feature based opinion summarization and visualization. arXiv preprint arXiv: 1504. 03068.

[5] Jayakody, J. P. U. S. D., & Kumara, B. T. G. S. (2021, December). Sentiment analysis on product reviews on twitter using Machine Learning Approaches. In 2021 International Conference on Decision Aid Sciences and Application (DASA) (pp. 1056-1061). IEEE.

[6] Chiorrini, A., Diamantini, C., Mircoli, A., & Potena, D. (2021, March). Emotion and sentiment analysis of tweets using BERT. In Edbt/icdt workshops (Vol. 3, pp. 1-7).

[7] Ouyang, X., Zhou, P., Li, C. H., & Liu, L. (2015, October). Sentiment analysis using convolutional neural network. In 2015 IEEE international conference on computer and information technology; ubiquitous computing and communications; dependable, autonomic and secure computing; pervasive intelligence and computing (pp. 2359-2364). IEEE.

[8] Amazon Product Reviews, Kaggle (2021).

https://www.kaggle.com/datasets/arhamrumi/amazon-product-reviews

[9] Hutto, C., & Gilbert, E. (2014, May). Vader: A parsimonious rule-based model for sentiment analysis of social media text. In Proceedings of the international AAAI conference on web and social media (Vol. 8, No. 1, pp. 216-225).

[10] Liu, Y. (2019). Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692.