Android Malware Classification Based on Fuzzy Hashing Visualization

Round 1

Reviewer 1 Report

This paper studies the feasibility of utilizing a deep learning model based on a CNN architecture for malware classification by converting a sample into a grayscale image composed of fuzzy hashes.

The paper reflects a relatively substantial work; however, there are significant issues in the proposed method. For more details, please refer to my comments on your submitted paper.

Below are some of my primary concerns:

·        The authors should clearly define:

o    Their input compared to the existing methods.

o   The applicability of their proposed method

To validate their proposed method, the authors must recenter their experiments around the existing methods and visualize the differences into graphs. They need to reorganize their experiments section to elaborate on their PoC.

·        The authors should explain:

o   How did they obtain the values presented in their tables?

o   The significance of their simulation graphs and how did they get them.

o   Elaborate on more simulation scenarios and graphs involving the other methods they compare their paper to.

Comments for author File: Comments.pdf

Please check the highlighted words on your document.

Author Response

Dear Reviewer,

Thank you for your valuable feedback. We have carefully reviewed your comments and have made the necessary revisions to the manuscript as per your suggestions. In the updated manuscript, you will find all the changes that address the concerns and recommendations outlined in the PDF.

We appreciate your time and effort in reviewing our work, and we believe that these improvements have significantly enhanced the quality of the manuscript. If you have any further comments or questions, please do not hesitate to reach out. Your feedback is precious to us, and we are committed to ensuring the highest quality for our research.

Thank you once again for your insightful review.

Sincerely, 

Horacio BAZAN et al.

Author Response File: Author Response.pdf

Reviewer 2 Report

Here are some review comments and suggestions for the research paper "Android Malware Classification Based on Fuzzy Hashing Visualization":

Abstract:

- The abstract provides a good high-level overview of the key points of the paper. Some minor suggestions:

  - Be more specific on the dataset used - mention the number of samples and families.

  - Quantify the improvement in accuracy over other methods.

Introduction: 

- Provides good background and motivation highlighting the threat of Android malware.

Related Concepts:

- This section gives helpful context on fuzzy hashing, APK structure, and NLP techniques used.

Related Work:

- Consider mentioning limitations of prior art to better highlight novelty of your approach.

Proposed Method:

- The two main stages are clearly explained - APK to image conversion and CNN classification.

- More details could be provided on the specific CNN architecture and training methodology.

Experiments:

- The experimental setup seems appropriate given the goals of evaluating the fuzzy hash images.

- More information could be included on metrics beyond accuracy used to evaluate performance. 

- Analysis of misclassifications may provide further insights.

Results: 

- Key results on accuracy with different image types and classes are summarized nicely in tables/graphs.

- Some further interpretation of the results would make this section more impactful.

Discussion:

- Good to compare accuracy of CNN versus SVM/KNN on same data.

- Limitations around imbalanced classes could be expanded on.

Conclusion:

- Effectively summarizes the main outcomes and contributions of the research.

- Consider adding some practical applications and future work based on this method.

Overall the paper is well-written and describes a novel approach for malware classification. My main suggestions would be to expand on details of methodology, provide more analysis of results, discuss limitations, and highlight practical implications. 

Author Response

Dear Reviewer,

Thank you for your valuable feedback. We have carefully reviewed your comments and have made the necessary revisions to the manuscript as per your suggestions. In the updated manuscript, you will find all the changes that address the concerns and recommendations.

We appreciate your time and effort in reviewing our work, and we believe that these improvements have significantly enhanced the quality of the manuscript. If you have any further comments or questions, please do not hesitate to reach out. Your feedback is highly valuable to us, and we are committed to ensuring the highest quality for our research.

Thank you once again for your insightful review.

Sincerely, 

Horacio BAZAN et al.

Author Response File: Author Response.pdf

Reviewer 3 Report

Review for “Android Malware Classification Based on Fuzzy Hashing Visualization”

This paper is well written. However, there are several suggestions that I would like to add in order to further enhance it:

1)     Please elaborate more on the introduction section regarding the context and current challenges. It seems rather detached from the following sections. You may try to better integrate introduction and provide more insights from the following sections.

2)     The list of references is broad, but some of them are published before 2020. Please include more novel references in the fields of ML such as:  https://doi.org/10.24846/v32i3y202303; https://doi.org/10.3390/sym14112304; https://doi.org/10.1016/j.engappai.2023.106030; https://doi.org/10.1016/j.cosrev.2022.100529

3)     The Figures are in good quality shape, however, please confirm the source of the figures throughout the manuscript.

4)     Findings are supported by the method and results. More limitations can be added, such as the method was applied on a single dataset.

5)     In the conclusion section, please provide more details to indicate and ensure the method replication.

Author Response

Dear Reviewer,

Thank you for your valuable feedback. We have carefully reviewed your comments and have made the necessary revisions to the manuscript as per your suggestions. In the updated manuscript, you will find all the changes that address the concerns and recommendations.

We appreciate your time and effort in reviewing our work, and we believe that these improvements have significantly enhanced the quality of the manuscript. If you have any further comments or questions, please do not hesitate to reach out. Your feedback is highly valuable to us, and we are committed to ensuring the highest quality for our research.

Thank you once again for your insightful review.

Sincerely, 

Horacio BAZAN et al.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Thank you for addressing most of my concerns.

One minor issue: the text in some of your diagrams and graphs is still illegible.

A moderate editing of the English language used in this article is required.

Author Response

Dear Reviewer,

Thank you for your valuable feedback. We have carefully reviewed the diagrams and graphs, and we agreed that improvements were needed. Specifically, we focused on Figures 1 to 5, where we increased the font size to enhance legibility and clarity. Additionally, we conducted a thorough review of the entire document and made necessary updates to address language-related issues. We believe these changes will significantly improve the overall quality of the paper.

Best regards,

Horacio Rodriguez-Bazan

Reviewer 3 Report

In the conclusion section, please provide more details to indicate and ensure the method replication.

Author Response

Dear Reviewer,

We appreciate your thoughtful feedback. After a thorough review, we have addressed your concerns by enhancing the conclusion section to include more detailed information. Our aim is to provide a comprehensive and replicable framework for future studies. We believe that the updated conclusion will contribute to the clarity and replicability of our research.

Additionally, we have conducted a careful review of the entire document and made necessary updates to address language-related issues. We are confident that these changes will significantly improve the overall quality of the paper.

Thank you for your valuable insights.

Best regards,

Horacio Rodriguez-Bazan