A Multi-Stance Detection Method by Fusing Sentiment Features

Round 1

Reviewer 1 Report

The paper outlines a methodology for detecting five stance categories using text sentiment features. While the document is well-crafted, it would benefit from some minor editing for English language usage. Consequently, the entire document should be reviewed to ensure a comprehensive understanding. 

The author uses the terms "method" and "model" interchangeably throughout the document. It is important to review and clarify these concepts to ensure a complete understanding of the methodology. Similarly, the author uses the terms "sentiment" and "polarity" interchangeably, which should be reviewed, clarified, or justified as necessary. 

The Abstract and Introduction sections should provide a clear context, problem statement, and details of the solution to engage the reader and ensure clarity. For example, the document would benefit from elaborating on the "significant influence" of stance on government policies or market strategies. Additionally, providing more details about the five categories of stance to detect would keep the reader interested and motivated to continue reading. 

While the first stage of the multi-stance indicator system is well-described, the extraction of sentiment features requires further clarification. It would be helpful to explain the role of stop words and who is responsible for classifying words such as adverbs. 

Section 5 introduces public information management, which would be better placed in the Abstract and Introduction. The document would also benefit from more coherence between Sections 4 and 5, which should relate directly to the Summary, Introduction, and Conclusions. If Section 5 is intended as a case study, it should be justified and explained more thoroughly, with consideration given to including other application scenarios.

Finally, the conclusions could be more robust and directly related to the experimentation, scientific contribution, and proposed solution methodology.

While the document is well-crafted, it would benefit from some minor editing for English language usage. Consequently, the entire document should be reviewed to ensure comprehensive understanding.

Author Response

Dear Reviewer,

We would like to express our gratitude for the time and effort you have dedicated to reviewing our manuscript titled "A multi-stance detection method by fusing sentiment features". Your insightful comments and suggestions have been invaluable in enhancing the quality of our work. We have carefully considered each point and have made the following responses and improvements, as detailed in the attached document.

We believe these changes have significantly enhanced the manuscript and addressed the concerns raised.

Thank you once again for your time and valuable insights.

Sincerely,
Weidong Huang, Jinyuan Yang
[email protected]

Author Response File: Author Response.pdf

Reviewer 2 Report

This is a paper on stance detection as a classification task. In general, the paper is well written, clearly structured and proposes a method showing promising results.

I think the introduction would benefit from more detailed discussion of possible applications of stance detection following this statement: line 24 stance information is of tremendous research significance.

The content of the sentiment lexicon (in section 3.2.1) needs more elaboration, what it includes (e.g., single words, multiword expressions, etc.), what are the stopwords, etc. 

In 3.2.2 the sentiment scores also need explaining how they are determined, or citing a source if they are taken from a particular study.

In presenting the method in 3.3 some references would be useful on how the equations are constructed and what these values show. Otherwise, the section is very informative and up to the point. Some are not standard in the field. In contract, extensive explanations on precision, recall and F1 score (heavily used across all areas of computational linguistics) in section 4.2 

In general, more details are needed on the text analysis, annotation and keyword extraction methodology. Otherwise, Table 8 in the last section is not clear and the reported results cannot be understood. More explanations are needed on the results as well.

Further, the stance categories are based on degree - strong/weak support, neutral, weak/strong opposition. However, in the last section the following categories are given: SS believes that the fault lies with the beaten child and backs the doctor's conduct in defending his rights. WS shows solidarity with the doctor by opposing the guardian of the beaten child, etc. SO believes that using violence to uphold rights was utterly unacceptable, while WO believes that the doctor just acted impulsively. In addition, Neutral discusses the details of the event, including the cause and effect of the incident, etc.

These categories are complex and include several events that should be evaluated independently. For example, an opinion can strongly support that fault lies with the beaten child but strongly oppose the doctor's conduct.

Minor notes:

Formatting: Some references are formatted as superscripts.

line 20 which have resulted > which has resulted

table 8 headings Keyword > Keywords

Author Response

Dear Reviewer,

We are writing to extend our heartfelt appreciation for the meticulous review of our manuscript, "A Multi-Stance Detection Method by Fusing Sentiment Features." Your expert insights and constructive suggestions have been instrumental in refining our research and elevating the manuscript to a higher standard.

After thorough deliberation on your feedback, we have made comprehensive responses and revisions, which are outlined in the attached document. We have taken great care to ensure that each of your points has been thoughtfully addressed and that the resulting changes have not only improved the clarity and coherence of our work but also strengthened the underlying research.

We are confident that these revisions have meaningfully addressed the concerns you raised and have significantly contributed to the overall integrity and impact of our manuscript.

We are immensely grateful for the opportunity to refine our work under your guidance and hope that the manuscript now meets the esteemed standards of the journal.

Thank you once again for your time, expertise, and the valuable insights you have provided.

Sincerely,

Weidong Huang, Jinyuan Yang

[email protected]

Author Response File: Author Response.pdf

Reviewer 3 Report

This research proposes a multi-stance detection method by fusing sentiment features. With a more comprehensive classification of positions, BACF can accurately capture users' stances and attitudes, recognize the differences and reasons between different stances, and guide public opinion management. This method is expected to be employed in public opinion, government policy, and market strategy. Although the method proposed in this paper achieves improved results in multistate detection, it relies on manual judgment when categorizing topics into different polarities and degrees of stance. Much research has been published worldwide on the multi-stance detection and similar methods. The paper is written in style, more like a lecture rather than a research article. The global innovativeness in research development has yet to be presented. Some figures and tables that involve novel worldwide research should be described and discussed in more detail to emphasize the paper's novelty. Please use this for the newest (2020-2024) Web of Science journal papers.

Author Response

Dear Reviewer,

Thank you for your thorough review of our manuscript, "A Multi-Stance Detection Method by Fusing Sentiment Features." We have carefully considered your comments and made the following revisions and responses, as detailed in the attached document.

We are immensely grateful for the opportunity to refine our work under your guidance and hope that the manuscript now meets the esteemed standards of the journal. We are confident that these changes have improved our manuscript and addressed your concerns.

Appreciate your valuable feedback and the time you have invested in this process.

Best regards,

Weidong Huang, Jinyuan Yang

[email protected]

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The comments and suggestions were addressed by the authors.

Reviewer 3 Report

In the future, writing articles should pay much more attention to global innovations. Emphasis should be placed on what you have done globally that is better than others.