Next Article in Journal
Correction: Li et al. An Efficient Multi-Level 2D DWT Architecture for Parallel Tile Block Processing with Integrated Quantization Modules. Electronics 2024, 13, 4668
Previous Article in Journal
AI-Assisted Inheritance of Qinghua Porcelain Cultural Genes and Sustainable Design Using Low-Rank Adaptation and Stable Diffusion
Previous Article in Special Issue
A Unified Model for Chinese Cyber Threat Intelligence Flat Entity and Nested Entity Recognition
 
 
Article
Peer-Review Record

Cost-Efficient Hybrid Filter-Based Parameter Selection Scheme for Intrusion Detection System in IoT

Electronics 2025, 14(4), 726; https://doi.org/10.3390/electronics14040726
by Gabriel Chukwunonso Amaizu 1,2, Akshita Maradapu Vera Venkata Sai 1,*, Madhuri Siddula 3 and Dong-Seong Kim 2
Reviewer 1:
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Electronics 2025, 14(4), 726; https://doi.org/10.3390/electronics14040726
Submission received: 31 December 2024 / Revised: 31 January 2025 / Accepted: 10 February 2025 / Published: 13 February 2025
(This article belongs to the Special Issue New Challenges in Cyber Security)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear authors,

In the next paragraphs my opinion about your manuscript

The paper is well organized with a clear introduction, detailed methodology and well-presented results. It proposes an innovative hybrid parameter selection scheme for IoT intrusion detection systems (IDS), with an impact on reducing computational costs and improving accuracy. Quantitative results show significant improvements in accuracy (98.12% for Dataset A and 96.45% for Dataset B).

The focus on the use of Gabor filters and deep learning techniques is relevant to intrusion detection problems in IoT environments, an area with growing challenges.

 

Aspects for improvement:

Although extensive, the review does not sufficiently contextualize the limitations of previous approaches, lacking more recent references from 2023/2024.

The description of some steps, such as the transformation from non-image to image data, could be more detailed to allow replication.

The results are based on public datasets (CSE-CIC-IDS2018 and ISCX-IDS-2012), but there are no tests or validation in real IoT scenarios, which may limit practical applicability.

They should explain figures 2 to 4.

I suggest more experiments with other datasets to validate the generalization of the proposed method.

Include more comparison details between the proposed model and other leading methods in the field, especially in terms of computational complexity.

Improve the explanation of the choice of some specific parameters (e.g., the values of the Gabor filter) and the hyperparameters of the base models.

Discuss more clearly the limitations of the proposed method and possible ways to overcome them.

Increase the resolution of some figures for better visualization, especially heatmaps and correlation graphs.

The results presented demonstrate significant improvements in accuracy and computational efficiency compared to conventional approaches. They contribute to solving practical problems in IoT environments, where intrusion detection is critical. They use modern techniques (e.g., CNNs and Gabor filters) in line with the state of the art. However, the lack of validation in real scenarios and the reliance on public datasets are limitations that reduce the direct practical impact.

Author Response

The comments are addressed in the Word document below.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The paper a structured four-phase architecture, addressing preprocessing, parameter selection, data transformation, and classification, which provides a systematic approach to intrusion detection in IoT environments. Also, the use of a hybrid filter-based parameter selection approach effectively reduces computational costs while maintaining high accuracy by eliminating redundant and highly correlated features.
Here are some thoughts to improve it:

 

The process of transforming non-image data into images and applying Gabor filters introduces significant computational overhead, which could hinder real-time applications in resource-constrained IoT devices.

What would be the downside of this approach? The scalability of the framework to larger and more diverse IoT networks has not been validated, which is critical for practical deployment in real-world scenarios.

 

 

Author Response

The comments are addressed in the Word document below.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

(1)     The framework and logical structure still need further refinement and adjustment. Related optimizations are also needed for the figures in the paper (such as figures 1, 5, 6, 9).

(2)     Authors should read the paper carefully to catch all the grammar and writing errors. For example: (1), (4), and (5) have an extra comma“.”.

(3)     How does the proposed method compare in terms of computational efficiency and detection accuracy against existing state-of-the-art IDS solutions, particularly in large-scale IoT environments? 

(4)     How does the use of image processing techniques contribute to the overall performance of the IDS?

 

(5)     What metrics were used to evaluate the detection accuracy and computational efficiency of the proposed system?

Author Response

The comments are addressed in the Word document below 

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

No

Back to TopTop