Detecting IoT Anomalies Using Fuzzy Subspace Clustering Algorithms
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors1) Any abbreviation must be explained before its first use or even at its first appearance in the work. Example: IoT = Internet of Things.
2) Some concrete results obtained by the proposed method should appear briefly at the end of the abstract.
3) The grammar of the English language must be checked, because there are many errors of expression. Example sentence of rows 31-33: "There is a huge applications comprising sensors that provide critical data that evolves over time, mostly as a result of the development of the IoT [1] along with their sources of real data generation."
a) You may say "a huge application" or "huge applications".
b) Elimines "over time"."
4) Use the academic expressions (with it is doing; not with we are doing; use the third person singular in the expressions instead of the first person). Example: As a result, "we are witnessing" a rapid surge of streaming and time-series data availability (row 33).
5) Instead of "Analysing such data can yield insightful information", use "Analyzing such data can yield insightful information" (row 34).
6) Citation [2] is missing.
7) The text in lines 86-89 has a smaller size.
8) Two equations numbered 6 appear.
9) Equation 12 is way too big.
10) Figures 1-5 require detailed explanations.
11) Figures 6-18 require individual detailed explanations.
12) Two Figures marked with 11 appear.
13) Tables 1-4 require individual detailed explanations.
14) Section 5 must be completely rewritten because it is difficult to follow in this nebulous form.
15) All the novelties must be highlighted and the results obtained with their help compared with those obtained by the known classical methods.
Extensive editing of the English language required
Author Response
Reviewer1.
Comments and Suggestions for Authors
1) Any abbreviation must be explained before its first use or even at its first appearance in the work. Example: IoT = Internet of Things.
Response1. All the abbreviations are properly explained at their first appearance in the paper as suggested by the honourable reviewer.
2) Some concrete results obtained by the proposed method should appear briefly at the end of the abstract.
Response2. Some concrete obtained results are incorporated at the end of the abstract of the paper as suggested by the honourable reviewer.
3) The grammar of the English language must be checked, because there are many errors of expression. Example sentence of rows 31-33: "There is a huge applications comprising sensors that provide critical data that evolves over time, mostly as a result of the development of the IoT [1] along with their sources of real data generation."
Response3. The paper has been properly checked. All errors including mentioned in the comments have been corrected. Sentences are corrected properly as suggested by the learned reviewer.
a) You may say "a huge application" or "huge applications".
Response3(a). The authors have incorporated the above as proposed by the learned reviewer.
b) Elimines "over time"."
Response3(b). The authors have eliminated the word, “over time”.
4) Use the academic expressions (with it is doing; not with we are doing; use the third person singular in the expressions instead of the first person). Example: As a result, "we are witnessing" a rapid surge of streaming and time-series data availability (row 33).
Response4. The academic expressions are used instead of the first person as suggested by the learned reviewer.
5) Instead of "Analysing such data can yield insightful information", use "Analyzing such data can yield insightful information" (row 34).
Response5. The corrections have been made accordingly as suggested by the learned reviewer.
6) Citation [2] is missing.
Response6.The citation [2] has been properly as citation [16] as pointed out by the learned reviewer.
7) The text in lines 86-89 has a smaller size.
Response7. The corrections have been made as pointed out by the learned reviewer.
8) Two equations numbered 6 appear.
Response8. The equations numbers are rearranged properly as pointed out by the learned reviewer
9) Equation 12 is way too big.
Response9. Equation 12 is written from existing literature cannot be made smaller.
10) Figures 1-5 require detailed explanations.
Response10. Figures 1-5 are explained properly as suggested by the learned reviewer.
11) Figures 6-18 require individual detailed explanations.
Response11. Figures 6-19 are individually explained as suggested by the learned reviewer.
12) Two Figures marked with 11 appear.
Response12. Two Figures marked as 11 are rearranged properly as pointed out by the learned reviewer
13) Tables 1-4 require individual detailed explanations.
Response13. Explanations of Tables 1-4 are given in the text as suggested by the learned reviewer.
14) Section 5 must be completely rewritten because it is difficult to follow in this nebulous form.
Response14. Section 5 has been re-written as per suggestions of learned reviewer. It has been broken into two subsections namely 5.1 Experimental Analysis and Results, and 5.2 Discussions.
15) All the novelties must be highlighted and the results obtained with their help compared with those obtained by the known classical methods.
Response15. All the novelties of our approaches are highlighted and comparative analysis has been done not only classical (traditional) fuzzy clustering approaches but also among the proposed approaches.
Comments on the Quality of English Language: Extensive editing of the English language required
Response. The article has been thoroughly revised and extensive editing of the English language has been done.
All the changes in the text are highlighted using Track changes.
We appreciate the honourable reviewer for his valuable comments and suggestions and will incorporate more if requires.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsIn this article, mixed approaches consisting of nano topology and fuzzy clustering techniques are proposed for anomaly detection. Experimentally, it has been found that the proposed approaches outperform the aforesaid algorithms in terms of detection rates, accuracy rates, false alarm rates and computational times. Some suggestions are provided as follows.
1. Some word should be corrected. “ Eucidean distance” in page 6.
2. In the article, “table 1” should be corrected as Table 1, and “figure 1” should be corrected as Figure 1.
The authors proposed some clustering algorithm including FCM, GK, … , and NT-CM-FCM are performed the comparative analysis in Figures 13-18. The authors propose the comparisons to analyze the relations Dimension of the dataset with detection rate, accuracy rate, false alarm rate. It can be found that the comparative results are linear relationship before 20 Dimension of the dataset shown in Figure 13. However, the comparative results are another linear relationship after 20 Dimension of the dataset. In the other comparative results in Figures 14-18, one also can find the similar results with two kinds of linear responses. The authors can give more discussions and expressions in Section 5.
Comments on the Quality of English LanguageMinor editing of English language required
Author Response
Reviewer2
Comments and Suggestions for Authors
In this article, mixed approaches consisting of nano topology and fuzzy clustering techniques are proposed for anomaly detection. Experimentally, it has been found that the proposed approaches outperform the aforesaid algorithms in terms of detection rates, accuracy rates, false alarm rates and computational times. Some suggestions are provided as follows.
- Some word should be corrected. “Eucidean distance” in page 6.
Response1. The authors have checked the text properly and corrected all the typographic and grammatical mistakes as suggested by the learned reviewer.
- In the article, “table 1” should be corrected as Table 1, and “figure 1” should be corrected as Figure 1.
Response2. The authors have checked the text properly and corrected all the mistakes like figure, table etc as suggested by the learned reviewer.
- The authors proposed some clustering algorithm including FCM, GK, … , and NT-CM-FCM are performed the comparative analysis in Figures 13-18. The authors propose the comparisons to analyze the relations Dimension of the dataset with detection rate, accuracy rate, false alarm rate. It can be found that the comparative results are linear relationship before 20 Dimension of the dataset shown in Figure 13. However, the comparative results are another linear relationship after 20 Dimension of the dataset. In the other comparative results in Figures 14-18, one also can find the similar results with two kinds of linear responses. The authors can give more discussions and expressions in Section 5.
Response3. The authors appreciate the points highlighted by the reviewers. In Figure 13 (now 14), the relationship of detection rates with dimensions of KDDCup’99 dataset presented. It has been found that all the five traditional fuzzy clustering algorithms namely FCM, GK, GG, M-FCM, CM-FCM perform in terms of detection rates very well in the lower dimensional spaces, but performances decrease rapidly as the dimension increases. The rate of decrease very fast initially, though it appears little better later. However, the proposed algorithms specially NT-CM-FCM performs uniformly. Also in case other NT-based approaches the results are more or less uniform. That means the NT-based approaches follows one linear relationship with the dimension of the KDDCup’99 datasets. In case of Kitsune dataset similar observations can be made. From 14-18 (now 15-19), similar findings are observed i.e. the NT-based approaches maintain consistency with the dimensions of both the dataset in all other performance parameters like accuracy rate, false alarm rate etc. However, the traditional approaches are non-uniform, they show multiple linear relationship. The NT-CM-FCM fuzzy clustering algorithm shows the most uniform performance in all the given parameters. Section 5 is revised as per suggestion of the learned reviewer and the points have been incorporated.
Comments on the Quality of English Language: Minor editing of English language required
Response. The article has been thoroughly revised and extensive editing of the English language has been done.
All the changes in the text are highlighted using Track changes.
We appreciate the honourable reviewer for his valuable comments and suggestions and will incorporate more if requires.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe article describes the method proposed by the authors, which is a two-step hybrid approach consisting of subspace generation and fuzzy clustering.
Identification of subspaces is obtained by generating subspaces based on nano topology, and a comparative analysis is carried out on known approaches to fuzzy clustering leading to soft clusters.
The material presented is interesting and has great scientific potential. However, I suggest modifying the introduction section to emphasize the problem the authors are trying to solve and define it explicitly so that the reader does not have to guess.
The manuscript also lacks comments and discussion of the results directly below the presentation of them in the figures.
I also suggest revising the abstract to fully reflect and indicate the elements of novelty in the paper.
Are the results obtained only simulations? It would be good to present results based on an experiment other than simulation tests.
The summary should be more condensed and clearer.
Author Response
Reviewer3
Comments and Suggestions for Authors
The article describes the method proposed by the authors, which is a two-step hybrid approach consisting of subspace generation and fuzzy clustering.
Identification of subspaces is obtained by generating subspaces based on nano topology, and a comparative analysis is carried out on known approaches to fuzzy clustering leading to soft clusters.
The material presented is interesting and has great scientific potential. However, I suggest modifying the introduction section to emphasize the problem the authors are trying to solve and define it explicitly so that the reader does not have to guess.
Response. The Introduction section is revised to emphasize the problem as suggested by the learned reviewer.
The manuscript also lacks comments and discussion of the results directly below the presentation of them in the figures.
Response: The explanations of each Figure and Table is given properly and a separate subsection is introduced as 5.2 Discussions and all the comments, observations, analysis and discussions are given in detail in this subsection as suggested by honourable reviewer.
I also suggest revising the abstract to fully reflect and indicate the elements of novelty in the paper.
Response: The abstract is revised accordingly to reflect and indicate the elements of novelty as proposed by the learned reviewer.
Are the results obtained only simulations? It would be good to present results based on an experiment other than simulation tests.
Response: The results are based on actual experiments. There are multiple experiments with different features of the datasets and results are recorded and presented either tabulated form or using bar diagram or graphically.
The summary should be more condensed and clearer.
Response: The summary is made condensed and clear as suggested by the learned reviewer.
The changes in the text are highlighted using Track changes.
We appreciate the honourable reviewer for his valuable comments and suggestions and will incorporate more if requires.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsModerate editing of the English language required
Comments on the Quality of English LanguageModerate editing of the English language required
Author Response
The authors have made editing and corrected all the typo-graphic mistakes as suggested by learned reviewer. The authors convey their heartfelt thanks to the learned reviewer as his comments and suggestions have improved the authors work.
Reviewer 2 Report
Comments and Suggestions for AuthorsMy suggestions for this manuscript have been revised by the authors. It is acceptable for the current version.
Author Response
I, on behalf of all the authors would like to convey our heartfelt thanks to the respected editor and reviewer for taking the pain to review our article again. I would also like to incorporate any other changes or suggestions put forwarded by the respected reviewers in future.
Sincerely yours
Dr. Fokrul A. Mazarbhuiya
Reviewer 3 Report
Comments and Suggestions for AuthorsCongratulations on your good work. The changes made take into account my comments submitted previously. All I ask is that you review the article from a technical point of view and correct the standard of writing equations and the legibility of drawings.
Author Response
The authors have incorporated almost all the suggestions visible in the manuscript. I, on behalf of all the authors would like to convey our heartfelt thanks to the respected editor and reviewer for taking the pain to review our article again. I would also like to incorporate any other changes or suggestions put forwarded by the respected reviewers in future.
Sincerely yours
Dr. Fokrul Alom Mazarbhuiya
Corresponding authors
