PONE-D-23-39359SCRQE: Subjective Comparative Relation Quintuple Extraction from Questions in Product DomainPLOS ONE

Dear Dr. Fatemi,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please note that we have only been able to secure a single reviewer to assess your manuscript. We are issuing a decision on your manuscript at this point to prevent further delays in the evaluation of your manuscript. Please be aware that the editor who handles your revised manuscript might find it necessary to invite additional reviewers to assess this work once the revised manuscript is submitted. However, we will aim to proceed on the basis of this single review if possible. The reviewer has raised a range of concerns, including methodological clarity, the context and overall justification for this work compared to other work in the field, and appropriate comparisons with state-of-the-art reports on similar methods. Please ensure you address each of the reviewers' comments when revising your manuscript.

Please submit your revised manuscript by Apr 01 2024 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org . When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Hugh Cowley

Staff Editor

PLOS ONE

Journal Requirements:

1. When submitting your revision, we need you to address these additional requirements.

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at 

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and 

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. You indicated that ethical approval was not necessary for your study. We understand that the framework for ethical oversight requirements for studies of this type may differ depending on the setting and we would appreciate some further clarification regarding your research. Could you please provide further details on why your study is exempt from the need for approval and confirmation from your institutional review board or research ethics committee (e.g., in the form of a letter or email correspondence) that ethics review was not necessary for this study? Please include a copy of the correspondence as an "Other" file.

3. Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified (a) whether consent was informed and (b) what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If your study included minors, state whether you obtained consent from parents or guardians. If the need for consent was waived by the ethics committee, please include this information.

If you are reporting a retrospective study of medical records or archived samples, please ensure that you have discussed whether all data were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have data from their medical records used in research, please include this information.

4. In your Methods section, please include additional information about your dataset and ensure that you have included a statement specifying whether the collection and analysis method complied with the terms and conditions for the source of the data.

5. Please note that PLOS ONE has specific guidelines on code sharing for submissions in which author-generated code underpins the findings in the manuscript. In these cases, all author-generated code must be made available without restrictions upon publication of the work. Please review our guidelines at https://journals.plos.org/plosone/s/materials-and-software-sharing#loc-sharing-code and ensure that your code is shared in a way that follows best practice and facilitates reproducibility and reuse.

6. We note that your Data Availability Statement is currently as follows: All relevant data are within the manuscript and its Supporting Information files.

Please confirm at this time whether or not your submission contains all raw data required to replicate the results of your study. Authors must share the “minimal data set” for their submission. PLOS defines the minimal data set to consist of the data required to replicate all study findings reported in the article, as well as related metadata and methods (https://journals.plos.org/plosone/s/data-availability#loc-minimal-data-set-definition).

For example, authors should submit the following data:

- The values behind the means, standard deviations and other measures reported;

- The values used to build graphs;

- The points extracted from images for analysis.

Authors do not need to submit their entire data set if only a portion of the data was used in the reported study.

If your submission does not contain these data, please either upload them as Supporting Information files or deposit them to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of recommended repositories, please see https://journals.plos.org/plosone/s/recommended-repositories.

If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent. If data are owned by a third party, please indicate how others may request data access.

7. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information. 

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors present a new dataset, denoted by the acronym SCRQD, containing

1,275 subjective comparative questions from the smartphone domain. Inspired

by (Liu et al, 2021), the questions are translated into (sets of)

quintuples denoting subject entity, object entity, compared aspect,

constraint, and preference. Together with the SCRQD dataset, a new SCRQE

system for extracting these quintuples from the input questions is

presented and evaluated.

The SCRQE system is designed as a pipeline with three stages that are

evaluated separately as well as in combination. The performance is partly

compared to other similar approaches although the details of the comparison

are not always clear.

A major drawback of the text lies in exaggerative claims about the

uniqueness of the presented approach which then result in the fuzzy

comparison. The authors claim that "there is no publicly available dataset

to derive subjective comparative relations from the questions". This is

definitely not true as comparatives in NLP are studied for decades and

simple search reveals tens of works in this regard.

Moreover, the presented approach is not specific to questions at all

despite the authors stating the opposite. They do not link or offer linkage

to possible answers and the RoBERTa-based analysis does not use anything

different from standard comparative sentence analysis. Other (mostly)

ignored sources of comparison are numerous aspect identification datasets

and approaches, although some are briefly mentioned in the evaluation of

the first stage of the system.

The Related works section offers relatively good overview of the approaches

(although concentrating on methods from 2021 and older) but these

approaches are not compared to the proposed system and the dataset

and no explanation of possible differences is offered.

The decisions for specific annotations and labels are in some cases unclear

even with the provided definitions and examples. For example, why

"reasonable" (line 275) is Non-Gradable? Is it different from

"good/better"? The definitions of preference types (lines 532-580) are

insufficiently explained. What is "a comparison without a clear reference

point" (XOR) or "a degree of uncertainty" (X)? For example, what are the

differences in meaning between the following four different annotations

from the text:

Is it better or not to upgrade A to B? (X)

What is the satisfaction level of A compared to B? (X-Strong Better)

Which phone is better overall performance, A or B? (XOR-Better)

Does A have better quality than B? (Better)

What are the inter-annotator agreement values for these fuzzy categories?

From the stylistic point of view, the text is flooded with too many

acronyms, making the comprehension and navigation in some parts quite

difficult.

Another example of "exaggeration" can be seen in the emphasis on the

Multi-Task Learning (MTL) approach used in the system where the actual

difference is just between multiple binary classifiers (here single task

approaches) and one multi-class classifier. Multi-class token

classification is the most standard way of named entity recognition which

exactly corresponds to the underlying task of the presented system. Here

using three separate single-task or binary classifiers would be more

unusual.

The schema in Figure 2 of the model for the entity extraction task is

misleading, displaying three separate label outputs from the input

question. The correct form is a sequence-to-sequence process resulting in

token labels (one to one for each token in the question).

The description of the second phase is unnecessarily long (lines 364-500,

more than 5 pages). The text mostly repeats the same (relatively simple)

procedure several times. For example, Tables 8 and (later) 11 are not

justified as the other 3 methods are not used at all here. Algorithm 1 does

not offer anything new to the text. One page should be enough here.

What is left unexplained, is the distinction between Subject and Object

(Table 9) where in some sentences there are only multiple subjects and in

others subjects with objects without explanation.

Is the second phase actually needed? What if the differentiation between

subject and object was expressed by the BIO labels in the first phase?

Dropout of 0.0 in Table 20 is slightly unusual. Moreover, this parameter is

not included in the list in 4.2. Hyper-Parameter Optimization with other

values.

Table 24 with a Comparison of Aspect Extraction models ignores some works

with better results, e.g.

Wu, Z., & Ong, D. C. (2021). Context-guided bert for targeted

aspect-based sentiment analysis. In Proceedings of the AAAI conference

on artificial intelligence (Vol. 35, No. 16, pp. 14094-14102).

or even the work [24] from the text presenting BERT-pair which was,

however, not included in this comparison table.

The details of the comparison evaluation are missing. Did you run the

compared models yourself? What is e.g. the "classification over

single-sentence on the CompSent-19 dataset" compared in Figure 7? What is

the Different label in Figure 8?

Table 26 displays relatively low values for the None class accompanied by

high numbers for the other two classes. What are the misclassifications

then? A detailed confusion matrix here could help to clarify.

Errors in text:

line 50, "we fine-tuned the SCRQD dataset" -> dataset is not fine-tuned

line 54, "superiority over existing model" -> "models"

line 156, "REDDY and KMahesh" -> "Reddy and Mahesh Babu"

line 162, "et al. [11]" -> "Liu et al. [11]"

Table 6, "(Q2, 4-tuple2)" in Example 1 and "(Q1, 4-tuple1)" in Example 2

incorrectly refer to the other question

Table 7. with Illustration of BIO tags is completely mistaken leading to

entity "then iPhone", aspects "better" and "a", and constraints "Samsung

smartphone" and "12 with a price below 600".

line 382, "As highlighted by Sun et al. [23]" - [23] is Devlin et al.

line 576, "questions that assessing the intensity or degree" -> "that assess"

line 769, "p_i intersection g_i" - not all quintuple elements are sets, are they?

line 779, "The Binary matching strategy emerged as the most efficient"

- claiming that the most permissive evaluation is "efficient" is strange

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean? ). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy .

Reviewer #1: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/ . PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org . Please note that Supporting Information files do not need this step.

PONE-D-23-39359R1SCRQE: Subjective Comparative Relation Quintuple Extraction from Questions in Product DomainPLOS ONE

Dear Dr. Fatemi,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Jul 26 2024 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org . When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Leona Cilar Budler

Academic Editor

PLOS ONE

Additional Editor Comments:

There are some suggestions and comments from reviewers to improve your paper. Please take them into account to improve your paper.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: N/A

Reviewer #2: N/A

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have provided a substantial revision of the original text, adding

a number of previously missing details of the work. While the new text has

significantly improved the critical points of the authors' work, several

new questions arise accompanying the added details.

One of the crucial points lies in the subject difference to most existing

datasets, i.e. to working with comparative questions instead of comparative

sentences. The authors point at the differences (and complications) between

these two data approaches, but a motivation for processing questions is

missing. A possible reason could be connected with "answers", that are

natural companions of questions, but these are not touched at all in the

text. The problems can be seen also in the dataset preparation procedure

where the authors admit that "due to the scarcity of questions that exactly

matched our research criteria, we modified existing questions and

formulated new ones to diversify our dataset" (line 511).

This motivation also concerns the two main data format differences to

previous approaches, i.e. the subject-object distinction (or the CEI task)

and the detailed comparison classification (or the CPC task) to 14 categories

instead of just Better/Worse. The impact of these changes is not evaluated

in the text. The dataset statistics of the subject-object (in Reply to

reviewer, missing in text) reveal that the number of non-standard cases,

i.e. different from subject-subject, is less than 10%. The examples of the

detailed comparison categories are not always clear, e.g. why in "Why is

the iPhone 10 camera inferior compared to the iPhone XS" denotes "inferior"

a "Strong Worse" category (line 942)?

In line 272, the presented SCQRD dataset is said to enable "robust model

training and evaluations across different languages and domains". How? The

dataset is limited to English and to one domain.

One of the main novelties of the presented approach lies in exploitation of

the RoBERTa base model as published by Facebook research in 2019. Since

that time, a number of newer, larger and more capable models have been

published. It thus may be expected that simply interchanging the underlying

pretrained model can substantially improve the results. Have you

experimented with newer models?

The only results which offer true comparison with related works and

a previously published dataset are the numbers in the chart of Fig 12

"Analysis of the SCRE vs. others on Camera-COQE [5] across matching

strategies". The authors still do not present other works that have

published better results on the Camera-COQE dataset, e.g.

Yang, Zinong, et al. "UniCOQE: Unified comparative opinion quintuple

extraction as a set." Findings of the Association for Computational

Linguistics: ACL 2023. 2023.

with 31.95 of Exact F1 score vs 23.8 by the current approach in Fig 12.

Fig 10 and 11 both present analysis results on the SCQRD dataset, first

with detailed values for the SCRQE system, then comparing the results to

two other approaches. However, the numbers for SCRQE do not match in the

figures, e.g. the Binary values are 43.12, 60.76 and 50.16 for precision,

recall and F-score in Fig 10 but the displayed Binary score for SCRQE in

Fig 11 is 47.73. What does it denote?

The description of the Inter-Annotator Agreement (IAA) analysis is greatly

extended in the current revision, however, it also showed a major

discrepancy in the way how this measure was used. The main reason for

measuring IAA lies in offering an estimate of how "difficult" the annotated

(sub)tasks are for human annotators as an approximate comparison to the

system performance. IAA must thus be measured on independently annotated

datasets without any kind of post-annotation iterative processing,

discussions and refinements of the measured data. Here, the authors have

used IAA for a different purpose, i.e. for reducing the disagreement in the

final dataset. What is also still not clear from the text is whether each

of the three annotators has processed all 1275 dataset questions.

In lines 1324-1336, the authors explain imprecise usage of the intersection

operation applied to non-sets as a "metaphorical interpretation of

intersection". However, equations are to provide exactness to textual

descriptions, not metaphors. A correct way is to adjust the equations to be

mathematically sound. It may be seen that the flawed equations propagate

from the original paper (Liu et al, 2021), which is, however, not cited in

this section. A fixed interpretation of these equations may be found in

another follow-up paper

Q. Xu et al., "GCN-based End-to-End Model for Comparative Opinion

Quintuple Extraction," 2023 International Joint Conference on Neural

Networks (IJCNN), Gold Coast, Australia, 2023, pp. 1-6, doi:

10.1109/IJCNN54540.2023.10191436.

where the intersection is applied to the quintuples, not to their elements.

The main difference is that "g_i indicates the i−th gold answer quintuple,

p_i denotes the i−th predicted quintuple result" (Xu et al, 2023, correct)

versus "p_i and g_i represent the i-th element in the predicted and gold

quintuple comparative relations" (both Liu et al, 2021 and the current

article, incorrect).

In lines 107-109 and repeated further, the Fleiss' Kappa is expressed as

percentages, which does not correspond to the definition of this measure

which reaches values in the interval from -1 to 1.

Table 17 compares the element extraction task results for three variants of

the implementation with bold numbers for the MTL+adapter model even in

cases where there are better results for the other variants. Why?

Errors in text:

line 225, "RoBERTa, which is finely tuned" -> "... is fine-tuned"

line 743: ", where:" but nothing follows

line 845, wrong section number 3.2.2.1. after 3.3.2

lines 1081-1084, the same statement twice:

- "categories ... either leave the direction of preference or the roles

of the entities ambiguous or undefined."

- "In these categories, one or more of these elements—either the

direction of preference or the roles of the entities—are typically

left ambiguous or undefined."

Reviewer #2: Concerns are as follows:

1. The methodological level of the narrative in the abstract be shortened to focus more on the problem and innovation.

2. Innovations are too broad, and metrics and data sets alone are insufficient as a fundamental basis for innovation.

3. SCQRD is a redundancy about the methodology, and irrelevant content needs to be removed, e.g., datasets, traditional methods, performance metrics.

4. There is a need for an overall block diagram of the SCQRD rather than a textual representation of the components.

5. There is a need for updating references.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean? ). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy .

Reviewer #1: No

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/ . PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org . Please note that Supporting Information files do not need this step.

PONE-D-23-39359R2SCRQE: Subjective Comparative Relation Quintuple Extraction from Questions in Product DomainPLOS ONE

Dear Dr. Fatemi,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Feb 24 2025 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
• A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
• An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Leona Cilar Budler

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

Additional Editor Comments:

There are some minor issues listed that authors need to resolve to consider this paper for publication.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1:  The authors have again substantially extended the article text (56 pages vs

32 pages in version 1) with detailed reactions to reviewers comments and

further work of the authors. The presented SCQRD dataset has been expanded

almost three times offering its enlarged version denoted as

Smartphone-SCQRE and a version adapted from the public CompSent-19 dataset

named here as the Brands-CompSent-19-SCQRE dataset. This has led to

improved coverage of previously underrepresented question types. Both

datasets have undergone a newly adapted annotation process with more

annotators and improved Inter-Annotator Agreement evaluation.

In the previous round, the authors have based the model architecture on the

RoBERTa base model. In the current text, the proposed system exploits

a specific model of RoBERTa_base_go_emotions, which is mentioned more than

40 times in the text (even used as one of the article keywords) but not

properly introduced until Section 3.3.1 at page 22. Moreover, the citation

with the model is not a correct one, as it refers to the original RoBERTa

model (without fine-tuning for emotions). If the model choice is so

crucial, for the task, it should be briefly introduced near its first

mention.

The RoBERTa_base_go_emotions model is evaluated together with newer

generative models (GPT-3.5-turbo-0613, Llama 2 70B Chat, and Qwen 1.5 7B

Chat) that were employed using their native prompt-based tasks. However,

for the purpose of Text Classification (like RoBERTa_base_go_emotions), the

generative models can be standardly fine-tuned with extra classification

layer(s). This usually offers better results than just prompting. But even

evaluating the prompt-based approaches (with the level of details as

offered in the article) is a valuable experiment.

The presented SCRQE model has been newly trained with the extended datasets

and with augmented reversed comparative statements which improved the

generalization capabilities and the resulting scores of the proposed model

so that it now outperforms a previously better model.

Overall, I believe the article is acceptable, with the correction of the

late introduction of the RoBERTa_base_go_emotions model.

Reviewer #2:  (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean? ). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy .

Reviewer #1: No

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/ . PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org . Please note that Supporting Information files do not need this step.

SCRQE: Subjective Comparative Relation Quintuple Extraction from Questions in Product Domain

PONE-D-23-39359R3

Dear Dr. Fatemi,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager®  and clicking the ‘Update My Information' link at the top of the page. If you have any questions relating to publication charges, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Leona Cilar Budler

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

No further comments

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: In the preceding round, the only suggestions proposed were the introduction

of the RoBERTa_base_go_emotions model and a mention of the fine-tuning

approach for generative models as future work. Both of these suggestions

are addressed in the text, therefore, the article is considered acceptable.

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean? ). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy .

Reviewer #1: No

Reviewer #2: Yes:  weibing wan

**********