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.
In regards to the revised manuscript, I have checked it carefully, and all the proposed issues have been addressed well. Moreover, one reviewer also accepted it for publication. In this case, I would like to accept it now.
The authors have reported Literature references, and sufficient field background/context is provided. In addition, literature references are sufficiently provided.
The authors have described the experiment with sufficient details & information.
The authors have revised the findings part and included the figures.
Well done.
In the opinions of two reviewers and mine, this manuscript should undertake a major revision.
**PeerJ Staff Note:** Please ensure that all review, editorial, and staff comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
- The abstract is informative but could be tighter. Cut back on background details and lead with the core contributions, method, datasets, and headline results.
- In the introduction, make the gap more explicit: state what past work couldn’t do, and how TTGNet-AMD addresses that.
- The related work section would be clearer if split into subsections (e.g., static, dynamic, hybrid detection) with a quick comparison at the end to highlight novelty.
- Discuss potential biases in CICMalDroid2020 and Drebin (e.g., app categories, malware types, age of samples) and how they may affect results.
- Explain why you used feature-level fusion instead of decision-level or hybrid fusion approaches.
- Give the paper a careful English language check for grammar, clarity, and smoothness.
- Clarify why the model stacks LSTM, Transformer, and GCN in that order — why not another arrangement?
- Describe how you chose key hyperparameters and whether you explored alternatives.
- Show training and validation loss curves to reassure readers the model converges stably and avoids overfitting.
- Since Banking malware detection lags behind other categories, suggest model tunings or data augmentation to improve it.
- When claiming superiority over baselines, run statistical significance tests to back up the claim.
- Back up confusion matrix results with examples of misclassifications and what might cause them.
- Test alternative fusion strategies and measure the impact of each feature type individually.
- Train on one dataset, test on the other, and report results to evaluate real-world generalization.
- Discuss how TTGNet-AMD might fit into commercial antivirus tools or enterprise mobile security.
- Include a short “Limitations” section to be transparent about current weaknesses and future improvements.
- Incorporate a broader comparison with some very recent studies in the Android malware detection space, particularly those that focus on feature optimization and multimodal fusion strategies.
- Add a summary table in the related work section showing datasets, features, models, and performance of prior studies for easy comparison.
The manuscript uses clear, professional English throughout, but it requires additional editing.
Literature references do not provide enough field background or context.
The professional article structure, figures, and tables must be redesigned and enhanced.
Not described in the right way.
The findings are Valid.
1. What are the precise research gaps that this study addresses compared to prior work?
2. Can the authors explicitly connect the stated objectives with the experimental/theoretical results?
3. How do the findings compare quantitatively with existing literature?
4. What are the assumptions and constraints in the experimental/simulation setup, and how might they bias the outcomes?
5. How were parameters chosen (e.g., material properties, laser settings, simulation variables)?
6. Were sensitivity analyses conducted to check the robustness of the results?
7. What are the mechanisms influencing [specific laser/material/system behavior] under the chosen experimental conditions?
8. How can the system be optimized to improve performance, efficiency, or stability?
9. What role do key parameters (dispersion, nonlinearities, or operational settings) play in controlling the observed dynamics?
10. To what extent can this approach address current limitations in [application field] compared to conventional methods?
11. Please revise all figures and tables; all of them are unclear.
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.