Automatic Detection of Multilevel Communities: Scalable, Selective and Resolution-Limit-Free^†

Round 1

Reviewer 1 Report

Community detection is a hot topic in social network analysis. The paper proposes an improved method for community detection based on a scalable community “fitness function.”  The new idea not only performs excellently on large heterogeneous networks without being affected by the res-olution limit problem, but also detects more multilevel communities, including the intermediate levels, in deep hierarchical networks.  Both the  background and the motivations are well explained. The organization  is clear and the results are new and sound. So, the paper is acceptable.  Some related referencea are encouraged to be cited.  Gaolin Chen, Shuming Zhou, Min Li, Hong Zhang, Evaluation of community vulnerability based on communicability and structural dissimilarity, Physcia A 566(2021) 125598.

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear Author,

I have made a few observations after reading your article. Your article might be considered for publication after fixing the ensuing minor review comments. However, kindly correct the reviewer comments for your expectation. 

In this manuscript, the author wrote an article entitled “Automatic detection of multilevel communities: scalable, selective and resolution-limit-free” suitable for publication, but the concerned author has to rectify the below-mentioned minor review comments in the “Applied Sciences”.

I have made a few observations after reading this article. This article might be accepted for publicaton after fixing the ensuing minor review comments. However, ask the author to correct the following:

 

1. What makes detecting communities so crucial in societal networks? Which algorithm for identifying communities performs the best?

2. Just how does Louvain community detection function? In other words, does Louvain use machine learning?

3.  To better understand machine learning, what exactly is a heuristic algorithm? Can you name a few different kinds of heuristics?

4.  When comparing heuristics and algorithms, what are the key differences? Heuristic algorithms are defined in the context of artificial intelligence.

5. Where does one find the formula for normalised mutual information? What's the dissimilarity between mutual information and correlation?

6. In what ways can one actually use a neural network? I was wondering how many distinct categories of network applications exist.

7. Explain the concept of a community in the context of a social network graph. How does the Girvan Newman algorithm locate the various communities present in the graph?

8. What are the different types of community detection techniques that are available?

9. Which algorithm is responsible for the detection of hierarchical communities?

10. Which method is used to find communities that overlap with one another?

 

 

It would be best if you corrected the above comments and resubmitted them according to your expectations.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

In this paper, the authors propose an improved method for community detection based on scalable community fitness function. In particular, they introduce a new parameter that improves its scalability and a strict strategy to filter the output. 

The proposed method sounds interesting, However, I have the following comments that need to be addressed before accepting the paper:

1. More experiments should be conducted on larger networks to test the performance of the proposed approach as the network's size increases

 

2. Comparison with other state-of-the-art algorithms is not available. I suggest to compare the performance proposed method in detecting community structure with other methods using different quality metrics such as normalized mutual information, adjusted Rand index, etc...

 

3. The figures' resolution is low, they need to be improved.  

 

4. Does Formula3 apply to both binary and weighted networks?

 

 

Minor:

The abstract language need to be revised and the verbs tenses should be consistent

 

 

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf