Detection and Localization of False Data Injection Attacks in Smart Grids Applying an Interpretable Fuzzy Genetic Machine Learning/Data Mining Approach
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors
1. The paper mentions using MATPOWER for data generation but fails to specify crucial attack injection parameters like maximum perturbation amplitude (ε), compromising experiment reproducibility.
2. The absence of comparisons with graph neural network (GNN) baselines weakens the methodological contribution, as GNNs have become standard in smart grid security research.
3. All experiments are confined to the IEEE 14-bus system; validation on larger-scale grids (e.g., IEEE 300-bus) is necessary to demonstrate method generalizability.
Comments on the Quality of English Language
The English could be improved to more clearly express the research.
Author Response
Please see the attached PDF file.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for Authors
Some commnets are given to improve the paper quality.
- In Page 7, our goal is to design a collection of 19 independent FR-BCs indicating, in a transparent and accurate way whether, in a given measurement point, we deal with FDIA (class label "Attack") or with normal operation of the smart grid (class label "Normal"). How to define transparent and accurate. Can these indicators accurately reflect the performance of the algorithm?
- How to determine the numbers of active fuzzy sets?
- The generated fuzzy rules are still too simple, and these rules can be set through experience. The innovation of this paper needs further discussion.
- The readability of this paper is poor, and the structure of the paper needs further optimization.
Comments on the Quality of English Language
The English could be improved to more clearly express the research.
Author Response
Please see the attached PDF file.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for Authors
This is an interesting and relevant article. It is a continuation of the research started by the authors in [6]. The main contribution of the authors of this article is the development of an effective machine learning and data mining (ML/DM) approach for the detection and localisation of false data entry attacks (FDIA) in smart grids. The authors present a set of fuzzy rule-based classifiers (FR-BC) that combine accuracy and transparency through a genetically optimised trade-off. The authors of the paper used an improved multi-objective evolutionary algorithm (M-OEOA), in particular, their generalisation of the SPEA2 algorithm is the SPEA3 algorithm, which demonstrates higher performance. The proposed approach outperforms 12 alternative methods in terms of interpretability and transparency, while remaining competitive in terms of accuracy. The authors also performed a detailed comparative analysis with other methods based on the open FDIA dataset. Therefore, this article is useful and provides new practical results.
The article makes a positive impression and contains significant information sufficient for publication. The problem under consideration is comprehensively covered. The authors of the article demonstrate a deep understanding of the relevant scientific literature and draw on a wide range of relevant sources in the field. The article cites key publications that are directly related to the topic of the study. The reference list is balanced and complements the general approach of the article. The reference list includes modern scientific works that provide an adequate theoretical basis.
The structure of the article is logical and well organised. The research results are clearly presented, analysed in detail and substantiated. The presentation of the material is clear and meets the academic standards of the English language. In the Conclusions, the authors compare the results of the proposed method with 12 alternative methods.
There are some remarks on the article:
1) No numerical results are displayed in the Abstract and Conclusions. The numerical results of this study should be added to the Abstract and Conclusions.
2) In the last two columns of Table 2, the authors provide numerical data on the accuracy of the measurement results, ranging from 63.0% on page 10 to 100% on page 16. However, the authors of the article did not pay enough attention to the analysis of the obtained numerical values of the measurement accuracy. The authors should provide more information on this.
3) What are the directions for future research based on the results of this study? In the Conclusions section, the authors of the article should indicate the prospects for further research.
The results of this article are important for practice and society. This article contributes to general knowledge. I suggest that the article can be accepted for publication with minor revisions.
Author Response
Please see the attached PDF file.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for Authors
No further comments
