Issue |
ITM Web Conf.
Volume 70, 2025
2024 2nd International Conference on Data Science, Advanced Algorithm and Intelligent Computing (DAI 2024)
|
|
---|---|---|
Article Number | 04019 | |
Number of page(s) | 9 | |
Section | AI and Advanced Applications | |
DOI | https://doi.org/10.1051/itmconf/20257004019 | |
Published online | 23 January 2025 |
Comparison of the Performance of Pegasos and Traditional Models in the Task of Sentiment Classification of Product Reviews
College of Information Engineering, China University of Geosciences, Beijing, China
Corresponding author: 1004211109email.cugb.edu.cn
Sentiment classification of a large number of commodity reviews is very important for customer selection and market trend prediction. The key to achieving high accuracy of sentiment classification is to select appropriate models for training. However, most of the existing research literature only uses traditional models or does not have a scientific comparison between models, and only provides training steps for a model that can be used for text sentiment analysis. Therefore, in this study, three models (two traditional models and one new model Pegasos) were selected for comparison. The same datasets and data preprocessing methods were selected for the three models to compare the final emotion classification accuracy, and finally, the experimental result with the highest accuracy of the Pegasos model was obtained. The study also analyzed the performance of different models in the experimental results and found the reasons for the poor performance of traditional models in handling sentiment classification tasks and the reasons for the good performance of Pegasos models. In the process of model hyperparameter tuning, the optimal number of iterations is 25 by comparing the classification accuracy under different iterations.
© The Authors, published by EDP Sciences, 2025
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.