All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
I am pleased to inform you that your work has now been accepted for publication in PeerJ Computer Science.
Thank you for submitting your work to this journal. I look forward to your continued contributions on behalf of the Editors of PeerJ Computer Science.
With kind regards,
[# PeerJ Staff Note - this decision was reviewed and approved by Sedat Akleylek, a PeerJ Section Editor covering this Section #]
The paper is generally clear and easy to understand. However, some more conventional wording and section naming might be helpful. For example, it is more conventional to call a section "Evaluation" or "Experiments" than "Test". Also, avoid using inaccurate expressions such as " training regimen", which sounds like the paper is in a medical context.
no comment
The paper claims that "...but also improves interpretability..." but shows no evidence elaborating or supporting that. I would suggest removing such claims, which might be distractive and confusing.
Reviewers have now commented on your paper. The reviewers have raised concerns regarding the problem statement, the experimental setup, and the results and discussions. Furthermore, the manuscript should be carefully revised by a fluent English speaker to improve readability. These issues require a minor revision. Please refer to the reviewers’ comments at the end of this letter; you will see that they advise you to revise your manuscript. If you are prepared to undertake the work required, I would be pleased to reconsider my decision. Please submit a list of changes or a rebuttal against each concern when you submit your revised manuscript.
Thank you for considering PeerJ Computer Science for the publication of your research.
With kind regards,
NO Comment
Methods are described with sufficient details
No Comment
This research presents the Multi-Scale Multi-Head Multi-Stage Network Intrusion Detection System (M³NID) was Proposed. The paper is interesting and such here are some minor comments for improving the article quality. The detailed comments are as follows,
The comment to the author are as follws,
1. Provide the key problems of the existing schemes and provide detailed problem statement with clear experimental results in the abstract section.
2. How does the proposed M³NID system differ from traditional IDS approaches?
3. The model is processed using three phases and write what are the main components used in three phases of M3NID and provide the role of each components.
4. The model uses PMSC for feature extraction and justify how it improve the detection accuracy and how it is better to standard models.
5. In the introduction provide the advantages and motivation for the development of the novel scheme or introduction section and then clearly sghowcase the contribution and novelty of the research.
6. The model employed dynamic gate fusion mechanism within PMSC module. Provide the functions of this mechanism.
7. The integration of BTA seems good and it is better to explain how this model enhances the temporal modelling and solves vanishing gradient issues.
8.Which datasets were used for testing, and why were they chosen?
9. Were ablation studies or sensitivity analyses performed to evaluate the individual contributions of PMSC, BTA, and the Transformer module?
10. How generalizable are the results to real-world, large-scale network traffic?
11. provide the practical application cghallenges of the model in conclusion section with future research direction.
12. provide how the model improve detection accuracy when considering specific types of attacks.
The writing is generally clear and gramatically correct. However, several typos and other writing issues can still be found and need fixing. For example, the line 110 to 112 are duplicates of line 107 to 109. The paper is self-contained with clear background and adequent references of intrusion detection and machine learning models such as CNN and Transformer. However, some figures, e.g., Fig 4, 6, 7, and 8 seem to have low resolution and are not of enough quality for printing, which require improvement.
The experimental design is generally reasonable and follows the practices of other works in the same field. However, in addition to the metircs reported in the paper, i.e., Acc (accuracy), DR (recall), and FPR, it would be easier to interpret the results to introduce a F-score, as it can better reflect the tradeoff between precision and recall.
The approach proposed by the paper has advantages on most datasets. However, for UNSW-NB15, while M3NID achieves higher Acc and DR, its FPR is relatively high, which is 91% higher than CNN-LSTM. Similarly, for the ablation study, while M3NID achieves the highest Acc and DR, its FPR is higher than PMSC-BTA. Providing a F-score or similar metrics that reflect the tradeoff between precision and recall can make the conlcusion more convincing.
The structure of the paper is well organized. But the introduction and background works must be given in details. The references from the recent years must be increased. The pictures and tables are readable.
There are very limited explanations about the data, test procedures, so on so forth. This sections must be improved.
The results and comparisons with the other recently proposed methods must be discussed.
The paper should be carefully revised by a fluent English speaker or a professional language editing service to improve the grammar and readability.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.