A Scientometric Analysis of Deep Learning Approaches for Detecting Fake News
Round 1
Reviewer 1 Report
Authors should make major changes before acceptance.
1. I strongly suggest authors abbreviate the abstract. One paragraph with no more than 250-300 words is better.
2. citation style may not be proper. For instance, line 121, [25][1] should be [1,25]. Line 151: [30-32].
3. please re-write section 1.4: some studies did… [37-39], and others did…[40-42]. Not the current one: in [37], it says…
4. proofreading is mandatory. Abundant grammar problems can be seen.
5. figure 22 is not clear to see the words. Figure 23 has overlaps of words.
6. describe what tools or software you use to make figure 24; what results can you get from this figure?
7. there is a huge problem: you just show me pictures but you don’t explain the contents and main results of the pictures. Readers may be new and fresh in this area; you need to tell them what we can learn from these figures. Please revise all the figures accordingly.
8. some references are not updated; you should update some new related references:
Highway Planning Trends: A Bibliometric Analysis
A Systematic Review of the Applications of Multi-Criteria Decision Aid Methods (1977-2022)
Bibliometric review of carbon neutrality with CiteSpace: evolution, trends, and framework
COVID Crisis and Tourism Sustainability: An Insightful Bibliometric Analysis
A Comprehensive Bibliometric Assessment on Software Testing (2016-2021)
Research Progress of Green Marketing in Sustainable Consumption based on CiteSpace Analysis
Author Response
- I strongly suggest authors abbreviate the abstract. One paragraph with no more than 250-300 words is better.
Response: We would like to thank the reviewer for the comment on the overall manuscript. The abstract has been kept within the suggested word count of 250–300. The changes have been highlighted for your consideration.
- citation style may not be proper. For instance, line 121, [25][1] should be [1,25]. Line 151: [30-32].
Response: We changed the citation style based on the feedback we received. Currently, [1,25] is used instead of [25] [1]. The modifications have been highlighted for your review.
- please re-write section 1.4: some studies did… [37-39], and others did…[40-42]. Not the current one: in [37], it says…
Response: As per your recommendation we changed the citation style. The phrase "authors in [42] examined" is now in use. The modifications have been highlighted.
- proofreading is mandatory. Abundant grammar problems can be seen.
Response:; In accordance with the reviewer's recommendation, proofreading has been done, and accordingly, the manuscript is revised. Thank you
- figure 22 is not clear to see the words. Figure 23 has overlaps of words.
Response: A revision has been made to recommended Figure 22 and 23. For your kind consideration, we have highlighted the captions of these images.
- describe what tools or software you use to make figure 24; what results can you get from this figure?
Response: With the recommendation of the reviewer, software has been described to make Figure 24 in that particular section. Results obtained from this depiction are also discussed. The amendments are highlighted for your kind consideration.
- there is a huge problem: you just show me pictures but you don’t explain the contents and main results of the pictures. Readers may be new and fresh in this area; you need to tell them what we can learn from these figures. Please revise all the figures accordingly.
Response: In accordance with your suggestions, to make it easier for readers to comprehend the content depicted in the given figures. We have added explanations to all the figures.
- some references are not updated; you should update some new related references:
Highway Planning Trends: A Bibliometric Analysis
A Systematic Review of the Applications of Multi-Criteria Decision Aid Methods (1977-2022)
Bibliometric review of carbon neutrality with CiteSpace: evolution, trends, and framework
COVID Crisis and Tourism Sustainability: An Insightful Bibliometric Analysis
A Comprehensive Bibliometric Assessment on Software Testing (2016-2021)
Research Progress of Green Marketing in Sustainable Consumption based on CiteSpace Analysis
Response: The manuscript now has updated references [71], [76], [83], and [95] that endorse the discussion, per the reviewer's advice. The changes are highlighted for your kind consideration.
Reviewer 2 Report
This review is about scientometric analysis of deep learning on detecting fake news. Works done before by other scientists are introduced. Also, based on 569 documents selected from Socups in recent 11 years, it gives analysis on intellectual and social structure. Using scientific mapping makes relationship between research constituents more visually and directly. And it aims to show new study prospects that requires further research. However, there are some issues need to be improved and considered by the authors.
First, In section 4,5 and 6, this part mainly introduces results. Authors present them as figures and graphs, therefore, for better presenting and understanding, specific analysis about figures and graphs is needed.
Second, in section 7, the final part concludes all the works done above. Authors need to draw more comprehensive and specific conclusion on all the results. Reveal deep connections and meanings among the results and summarize the trends of development. Then, the future scopes that needs further investigation can be acquired.
Third, words and expression are important for reading, there are a few typos and format errors need to be corrected.
For example, typos errors :In 3.2 “descripttive”, and in row 665“infolution”.
In the section of introduction, two sentences have the same meaning in row 81-86.
In the section of 3.2.1table 1, the row numbers are in the wrong position.
For row 353, 3.2.2 is mixed up with table1.
In figure 6 the third row’s typography needs improvement.
The format of titles needs to unify，like 3.2.4 and 3.2.5.
Figure 22’s picture is lack of sharpness.
In figure 23 ,words are overlapped.
Only after the issues above have been thoroughly thought and improved will this review be published.
Author Response
This review is about scientometric analysis of deep learning on detecting fake news. Works done before by other scientists are introduced. Also, based on 569 documents selected from Socups in recent 11 years, it gives analysis on intellectual and social structure. Using scientific mapping makes relationship between research constituents more visually and directly. And it aims to show new study prospects that requires further research. However, there are some issues need to be improved and considered by the authors.
First, In section 4,5 and 6, this part mainly introduces results. Authors present them as figures and graphs, therefore, for better presenting and understanding, specific analysis about figures and graphs is needed.
Response: We would like to thank the reviewer for providing valuable feedback. In accordance with the reviewer's recommendation, to make it easier for readers to better comprehend the content depicted in the given figures and graphs. We have added specific explanations to all figures. The modifications have been highlighted for your consideration.
Second, in section 7, the final part concludes all the works done above. Authors need to draw more comprehensive and specific conclusion on all the results. Reveal deep connections and meanings among the results and summarize the trends of development. Then, the future scopes that needs further investigation can be acquired.
Response: In order for future researchers to use this manuscript as a resource for additional research, we have modified the final section of it in accordance with the given suggestions.
Third, words and expression are important for reading, there are a few typos and format errors need to be corrected.
For example, typos errors :In 3.2 “descripttive”, and in row 665“infolution”.
In the section of introduction, two sentences have the same meaning in row 81-86.
In the section of 3.2.1table 1, the row numbers are in the wrong position.
For row 353, 3.2.2 is mixed up with table1.
In figure 6 the third row’s typography needs improvement.
The format of titles needs to unify，like 3.2.4 and 3.2.5.
Figure 22’s picture is lack of sharpness.
In figure 23 ,words are overlapped.
Response: Typos and format errors have been corrected in accordance with the reviewer's valuable feedback on the overall manuscript. Figures 22 and 23 have been revised. We have highlighted the captions for these images for your consideration.
Only after the issues above have been thoroughly thought and improved will this review be published.
Reviewer 3 Report
The paper is too long. Please focus on your area/topic of research. You need to state your scope of discussion of the paper. A literature review is not clearly conceptualised. Too long elaboration of methodology and findings report. I found some unnecessary figures and double data presented. Display and report necessary figures and data only, that are related to elaborate your research objective. The topic and subtopic of finding reports must refer to your research focus or objective. Moreover, Provide discussion, implication, and contribution of your research.
Author Response
There are visible formatting errors exhibited throughout the manuscript especially figure and table captions (i.e., page 9), inconsistent font styles, sizes (lines 160-210 v. lines 211 onwards), and spacing that needs more thorough proofreading, so it is readable for the audience.
Response: We appreciate the reviewer's comments on the overall manuscript. In accordance with the reviewer's recommendation, proofreading has been done, and accordingly, the manuscript is revised.
Additionally, the manuscript contains brazen misspellings (e.g., line 11) and sentence errors that make the messages and ideas being conveyed particularly difficult to comprehend (for instance, sections 1.2)
Response: Spelling errors have been corrected in accordance with the reviewer's recommendation. The title of Section 1.2 has also been updated so that it conveys the intended message. For your convenience, we have highlighted the captions for this section. Thank you.
First and foremost, the very term used as the topical application, “fake news detection,” is not even defined in any part of the article. Does fake news detection exclusively scope to technological advancements for this undertaking or fake news detection can be extended to the human ability of users to discern legitimate content from untruths (i.e., mis/disinformation recognition)? The authors correctly pointed out that fake news is an abstract term and is usually misinterpreted, but they used it repeatedly throughout the article, nonetheless.
Response: We appreciate the reviewer pointing out our mistakes. In response to the above-mentioned suggestions, the literature review section has been expanded to include fake news detection using deep learning techniques and advanced artificial intelligence techniques such as Explainable Artificial Intelligence (XAI) to explain the reasoning behind the model's decision.
Also, what is a non-information? Such a term is not included in the European Council’s Framework for Information Disorders. The authors did not provide any definition of this phenomenon, and they should include such. Information disorders typically only include mis/dis/malinformation.
Response: It was very helpful of the reviewer to point out our mistakes. The paper no longer contains "non-information." In the European Council's Framework for Information Disorders, misinformation, disinformation, and malinformation are types of information disorders. These have been included in our manuscript. Changes are highlighted for the reviewer’s kind consideration. Thank you.
Also, surprised that these analyses did not include IEEE Xplore, ACM’s DL, or even DBLP for their inclusion criteria, especially since deep learning, as a technical topic, is usually saturated in the Computer Sciences and tends to cluster in those unincluded libraries.
Response: We appreciate the reviewer pointing out our mistakes The analyses has included IEEE Xplore, ACM’s DL, and DBLP for their inclusion criteria.
Reviewer 4 Report
There are visible formatting errors exhibited throughout the manuscript especially figure and table captions (i.e., page 9), inconsistent font styles, sizes (lines 160-210 v. lines 211 onwards), and spacing that needs more thorough proofreading, so it is readable for the audience.
Additionally, the manuscript contains brazen misspellings (e.g., line 11) and sentence errors that make the messages and ideas being conveyed particularly difficult to comprehend (for instance, sections 1.2)
First and foremost, the very term used as the topical application, “fake news detection,” is not even defined in any part of the article. Does fake news detection exclusively scope to technological advancements for this undertaking or fake news detection can be extended to the human ability of users to discern legitimate content from untruths (i.e., mis/disinformation recognition)? The authors correctly pointed out that fake news is an abstract term and is usually misinterpreted, but they used it repeatedly throughout the article, nonetheless.
Also, what is a non-information? Such a term is not included in the European Council’s Framework for Information Disorders. The authors did not provide any definition of this phenomenon, and they should include such. Information disorders typically only include mis/dis/malinformation.
Also, surprised that these analyses did not include IEEE Xplore, ACM’s DL, or even DBLP for their inclusion criteria, especially since deep learning, as a technical topic, is usually saturated in the Computer Sciences and tends to cluster in those unincluded libraries.
I don’t see much value to this work aside from a mere weak and incomplete synthesis of the current and existing works on the topical application of fake news detection.
Author Response
We have received comments from three reviewers only.
Round 2
Reviewer 1 Report
it is worthy publication now.
Author Response
We would like to thank the reviewer for the comment on the overall manuscript.
Reviewer 2 Report
This vision of manuscript corrects the errors in the former one and adds more analysis. It also draws conclusions about the future development in deep learning of fake news detection. Those changes make this manuscript easier to understand and read.
However，there are still some format errors needs to be correct, like no spaces at the beginning of some paragraphs, errors in line 199 and table 7 and so on. So the typography of this manuscript needs a thorough check.
Only after all the issues above are addressed will the reviewer be accepted.
Author Response
Thank you; we appreciate the reviewer's feedback on the overall manuscript. In accordance
with the reviewer's recommendation, manuscript is revised.
Reviewer 3 Report
I don't find significant changes from the previous version. The contentment is not explored by flowing the title of the paper. It is difficult to understand the concept and the theory underlying the research question or objective. The paper is too long. The introduction started with very large issues and a lack of focus on the objective of the study. The findings are confusing, what research questions or objectives to answer? Too many findings but not specifically and clearly answer the research question.
Author Response
We appreciate the reviewer's pointing out our mistakes. The research questions have been
framed in such a way that the findings of the analysis can be related to them. We made every effort
to keep this study as short as possible. We believe that these revisions have resulted in a significantly
improved manuscript. Thank you.
Reviewer 4 Report
Include DBLP as part of the analysis
Author Response
As per given suggestions, DBLP has been included in this manuscript. For the kind
consideration of reviewer, track changes is used to mark changes to the manuscript
Round 3
Reviewer 3 Report
The revision had been completed. And congratulation.
Author Response
We would like to thank the reviewer for the comment on the overall manuscript.
