A Study on the Stability of Graph Edit Distance Heuristics
Round 1
Reviewer 1 Report
This manuscript studies the stability of the graph edit distance heuristics by defining relative error. The manuscript examines the number of solutions and edit cost ratios by an edit cost learning algorithm.
In my opinion, the manuscript seems to appear in the wrong place. There is no relation between the manuscript and electronics. I believe the journal of Electronics would like to see materials, energy bands, electrical carriers, etc. This manuscript is not suitable for Electronics, since its content mainly is the algorithm for computer science.
The manuscript was well written. However, one should be careful to say it is the first study of the heuristics stability for graph edit distance. Another suggestion is that the broader impact of the background is less discussed in the introduction.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
My correction with the manuscript are relatively minor, and are listed below.
It is suggested to the authors that the objectives of this study, the methods used to achieve them, and the final results be more clearly defined in the Abstract and Conclusion sections. The abstract explains the purpose of the work and includes the background information, but lacks a clear indication of the method used.
The introduction provides a good general background on the subject and gives the reader an idea of the wide range of possible applications of this technology. The methods used in this thesis are appropriate for the goal of the study.
In addition, corrections are needed in the presentation of the results, both in the standardization of font size, image resolutions and in the naming of images. Indeed, some images are impossible to interpret because of the titles they bear, especially because of the numerous abbreviations that have not been explained.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Title: A Study On The Stability of Graph Edit Distance Heuristics
This article seems well built and brings evidence of a phenomenon not yet fully understood and that certainly deserves further study.
Some minor points of revision are provided below
Abstract
I suggest including data in the results section to better clarify the results of your study
Introduction
Line 12-16: Needed a reference to justify this sentence
Method section
I suggest to including the statistical analysis in the same way the results in the results section (because in several part of this main document the authors wrote “Strong connection” but this point should be demonstrated with the data
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
The authors present an exciting paper where they propose the first formal study of the stability of GED heuristics, starting with defining a measure of these (in) stability, namely the relative error. Then, the effects of two critical factors on stability are examined, namely, the number of solutions and the ratio between edit costs. They provided general suggestions about these factors. Finally, they verify the relevance of stability to predicting GED heuristics performance by taking advantage of an edit cost learning algorithm. The manuscript’s results are reproducible based on the details given in the methods section. The manuscript is well written and should greatly interest the readers. In conclusion, they should mention more about their future work.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 5 Report
The authors have examined the initial formal study of the stability of GED heuristics, focusing on the relative error. The contribution is quite OK to me. The paper is quite well-written and easy to read and understand. In addition, the instability of a GED heuristic is quite well-defined. The experimental results and analyses on the GED stability are also quite convincing to me, especially the sizes of datasets (68 to 286). However, the relative errors of mbipartite and mIPFP on datasets Alkane and Acyclic are not so clear. It should be more elaborated in detail. Furthermore, some literature reviews about K-Nearest Neighbor Regression should be added and discussed more clearly. For example, the following papers should be mathematically added and discussed: A Sample-Rebalanced Outlier-Rejected k-Nearest Neighbor Regression Model for Short-Term Traffic Flow Forecasting. IEEE Access 8: 22686-22696 (2020), Color and Active Infrared Vision: Estimate Infrared Vision of Printed Color Using Bayesian Classifier and K-Nearest Neighbor Regression. PCM (1) 2015: 518-527, and Fault Detection and Isolation of a Pressurized Water Reactor Based on Neural Network and K-Nearest Neighbor. IEEE Access 10: 17113-17121 (2022). In conclusion, the paper could be accepted after proper revision. If not, I am afraid to reject this paper.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The authors have addressed my concerns about the topic and scope of the manuscript, and I recommend its publication.
Reviewer 3 Report
The main document was well improved
