Semantic-Based Public Opinion Analysis System

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Clarity and Depth of Methodology: The methodology section is comprehensive, covering data acquisition, sentiment analysis, and classification processes. However, it would benefit from more detailed explanations of how the sparse representation classification model was specifically adapted for sentiment analysis. Consider illustrating this with examples or a case study to enhance understanding.

Experimental Design: The experimental setup, particularly the choice of data sources and the analysis approach, is well-explained. However, the use of only one platform (PTT) may limit the generalizability of the results. It would be advantageous to include data from multiple platforms to validate the model's effectiveness across different types of online environments.

Results and Discussion: The results section provides a clear depiction of the system’s performance with specific examples. It is recommended to expand this section by discussing potential biases in the data or the system's limitations in handling diverse linguistic expressions or sarcasm, as noted.

Technical Improvements: The paper would benefit from a deeper dive into the technical challenges encountered, especially in processing and classifying diverse sentiments. More information on handling imbalanced datasets or rare sentiments would enrich the paper.

Future Work: The suggestions for future research are intriguing, particularly the focus on recognizing sarcasm and extending to speech synthesis. It would be beneficial to outline preliminary ideas on how these goals could be achieved or propose a roadmap for integrating these features into the current system.

Author Response

We would like to extend our appreciation to AE and the anonymous reviewers for their valuable and affirmative feedback. Please find enclosed a revised version of our paper, along with our comprehensive responses to each of the reviewers' comments, addressing each item and point in detail. We are confident we have carefully considered and addressed the reviewers' feedback and concerns. 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This paper introduces new system for semantic sentiment analysis. 

The paper is well-written and interest for a reader. All Figures and Tables are clear and visible. The structure of the proposed system is also well-written, which is important for a reader. The structure of the paper is also fine. 

The Authors might consider the following suggestions: 

1. Add more relevant and newer references in the paper (especially the Introduction). 

2. Why did you chose K-nearest neighbor algorithm, and did you consider some alternatives? 

3. Did you consider some other measures than Accuracy (such as Precision, Recall, ...), or is Accuracy enough?

4. Revise the sections at the end of the manuscript, such as Funding, Data Availability, Acknowledgments.  

Comments on the Quality of English Language

Quality of English Language is fine.

Author Response

We would like to extend our appreciation to AE and the anonymous reviewers for their valuable and affirmative feedback. Please find enclosed a revised version of our paper, along with our comprehensive responses to each of the reviewers' comments, addressing each item and point in detail. We are confident we have carefully considered and addressed the reviewers' feedback and concerns. 

Author Response File: Author Response.pdf